Final Project (server).ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "f1480de7-2744-4551-9e99-317f6b7893d2",
   "metadata": {},
   "source": [
    "# Imports and Function Definitions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "de881bb1-c0f2-436b-9f1d-ac9304afae45",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "\n",
    "from sklearn.model_selection import KFold, StratifiedKFold, train_test_split\n",
    "from sklearn.linear_model import ElasticNet\n",
    "from sklearn.ensemble import RandomForestRegressor, RandomForestClassifier, GradientBoostingRegressor, GradientBoostingClassifier\n",
    "from sklearn.metrics import mean_absolute_error, accuracy_score, mean_squared_error\n",
    "from sklearn.model_selection import TimeSeriesSplit, GridSearchCV, cross_val_score, KFold, StratifiedKFold\n",
    "\n",
    "from statsmodels.graphics.tsaplots import plot_acf, plot_pacf\n",
    "sns.set()\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')\n",
    "# from polygon import RESTClient\n",
    "\n",
    "import datetime as dt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "72c06572-8943-4f96-8a07-1e0696cb9c76",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "def tree_feat_importance(m, df):\n",
    "    return pd.DataFrame({'cols':df.columns, 'feat_imp':m.feature_importances_}\n",
    "                       ).sort_values('feat_imp', ascending=False)\n",
    "def ensemble_tree_feat_importance(m, df):\n",
    "    feature_importances = np.mean([\n",
    "    tree.feature_importances_ for tree in m.estimators_\n",
    "                                    ], axis=0)\n",
    "    return pd.DataFrame({'cols':df.columns, 'feat_imp':feature_importances}\n",
    "                       ).sort_values('feat_imp', ascending=False)\n",
    "def plot_fi(fi): return fi.plot('cols', 'feat_imp', 'barh', figsize=(15,20), legend=False)\n",
    "def xgb_feat_importance(m, df):\n",
    "    feature_importances = m.feature_importances_\n",
    "    return pd.DataFrame({'cols':df.columns, 'feat_imp':feature_importances}\n",
    "                       ).sort_values('feat_imp', ascending=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2abf351b-7c2e-4839-86b1-bd69644bb3e2",
   "metadata": {
    "tags": []
   },
   "source": [
    "# Cleaning and Merging"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 713,
   "id": "b4622c71-fae4-41f1-b807-d4e2f87a6c28",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-03-03T00:37:35.402169Z",
     "iopub.status.busy": "2023-03-03T00:37:35.402169Z",
     "iopub.status.idle": "2023-03-03T00:37:35.415168Z",
     "shell.execute_reply": "2023-03-03T00:37:35.415168Z",
     "shell.execute_reply.started": "2023-03-03T00:37:35.402169Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "client = RESTClient(api_key=\"18U85gkK04AWHrbE3oWvG9b3fF8JdlxQ\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 714,
   "id": "fa182899-8cc0-4bef-bb48-02dca8133d2c",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-03-03T00:37:35.643110Z",
     "iopub.status.busy": "2023-03-03T00:37:35.643110Z",
     "iopub.status.idle": "2023-03-03T00:37:38.524492Z",
     "shell.execute_reply": "2023-03-03T00:37:38.524492Z",
     "shell.execute_reply.started": "2023-03-03T00:37:35.643110Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "vxx = client.get_aggs(ticker='VXX', multiplier=5, timespan=\"minute\", from_=\"2022-04-01\", to=\"2022-09-01\", limit=50000)\n",
    "spy = client.get_aggs(ticker='SPY', multiplier=5, timespan=\"minute\", from_=\"2022-04-01\", to=\"2022-09-01\", limit=50000)\n",
    "uso = client.get_aggs(ticker='USO', multiplier=5, timespan=\"minute\", from_=\"2022-04-01\", to=\"2022-09-01\", limit=50000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 695,
   "id": "de3ad651-4930-4cf5-8c33-314e91c82686",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-03-03T00:33:07.022997Z",
     "iopub.status.busy": "2023-03-03T00:33:07.021997Z",
     "iopub.status.idle": "2023-03-03T00:33:07.037996Z",
     "shell.execute_reply": "2023-03-03T00:33:07.037077Z",
     "shell.execute_reply.started": "2023-03-03T00:33:07.022997Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "poly_dict = {'vxx':vxx, 'spy':spy, 'uso':uso}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 696,
   "id": "4255f238-0ff4-41b2-bae5-7af3bc7c12d4",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-03-03T00:33:07.039998Z",
     "iopub.status.busy": "2023-03-03T00:33:07.038998Z",
     "iopub.status.idle": "2023-03-03T00:33:07.633316Z",
     "shell.execute_reply": "2023-03-03T00:33:07.632257Z",
     "shell.execute_reply.started": "2023-03-03T00:33:07.039998Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "for data in poly_dict.keys():\n",
    "    poly_dict[data] = pd.DataFrame(poly_dict[data])[['close', 'timestamp']]\n",
    "    poly_dict[data] = poly_dict[data].rename(columns={'close':'Close', 'timestamp':'Local time'})\n",
    "    poly_dict[data]['Local time'] = pd.to_datetime(poly_dict[data]['Local time'], unit='ms')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 697,
   "id": "93fd99c8-09bb-4be1-ac11-95442ba7d6f4",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-03-03T00:33:07.634272Z",
     "iopub.status.busy": "2023-03-03T00:33:07.634272Z",
     "iopub.status.idle": "2023-03-03T00:33:07.664272Z",
     "shell.execute_reply": "2023-03-03T00:33:07.663282Z",
     "shell.execute_reply.started": "2023-03-03T00:33:07.634272Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "vxx = poly_dict['vxx']\n",
    "spy = poly_dict['spy']\n",
    "uso = poly_dict['uso']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 237,
   "id": "076bb000-a60d-4eb5-beae-b267e77688a6",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "fut = pd.read_csv('LIGHT.CMDUSD_Candlestick_5_M_BID_11.02.2020-11.02.2023.csv')\n",
    "gas = pd.read_csv('GAS.CMDUSD_Candlestick_5_M_BID_11.02.2020-11.02.2023.csv')\n",
    "btc = pd.read_csv('BTCUSD_Candlestick_5_M_BID_11.02.2020-11.02.2023.csv')\n",
    "bond = pd.read_csv('USTBOND.TRUSD_Candlestick_5_M_BID_11.02.2020-11.02.2023.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 238,
   "id": "c654650a-82e7-43a1-9e60-8b433b1be650",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "data_dict = {'fut':fut, 'gas':gas, #'vxx':vxx, 'spy':spy, #'uso':uso, \n",
    "            'btc':btc, 'bond':bond}"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "81caf05a-520b-4bab-ab89-cdab52b3a4a3",
   "metadata": {},
   "source": [
    "## Datetime and indexing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 239,
   "id": "f62b06f3-287f-48d7-ab11-5d0f1af2de1b",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "# idx = pd.to_datetime(data_dict['fut']['Local time'])\n",
    "# idx=idx.apply(lambda x: x.replace(tzinfo=None))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 240,
   "id": "cc3d83fb-f59e-4d95-ac6d-1510ceb057f6",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "for data in data_dict.keys():\n",
    "    if data in ['vxx', 'spy', 'uso']:\n",
    "        data_dict[data] = data_dict[data].set_index(idx)\n",
    "        data_dict[data] = data_dict[data].drop(['Open', 'High', 'Low', 'Volume', 'Local time'], axis=1)\n",
    "    else:     \n",
    "        ix = pd.to_datetime(data_dict[data][\"Local time\"].apply(lambda x: \" \".join(x.split()[0:2])), format='%d.%m.%Y %H:%M:%S.%f')\n",
    "        data_dict[data] = data_dict[data].set_index(ix)['Close']"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c11f3950-96a9-4578-995c-21734ad856a0",
   "metadata": {},
   "source": [
    "## Merging"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 241,
   "id": "205f196e-1ac1-4338-9a3b-d8b4c0452ec3",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "# df = pd.DataFrame(index=data_dict['vxx'].index)\n",
    "df = pd.DataFrame(index=data_dict['fut'].index)\n",
    "for data in data_dict.keys():\n",
    "    df = pd.merge(df, data_dict[data], left_index=True, right_index=True, suffixes=['', '_'+data], how='left')\n",
    "df = df.rename(columns={'Close':'Close_fut'})"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "580f11c0-18fe-45f5-b311-083ed2dcc708",
   "metadata": {},
   "source": [
    "## Cut to market hours"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 242,
   "id": "cfcfcb08-33a3-482c-b2e9-36a4ef68d44e",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "df['date'] = df.apply(lambda x: dt.datetime.strptime(str(x.name), '%Y-%m-%d %H:%M:%S').date(), axis=1)\n",
    "df['time'] = df.apply(lambda x: dt.datetime.strptime(str(x.name), '%Y-%m-%d %H:%M:%S').time(), axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 243,
   "id": "8c1355ae-d13e-4a70-947d-aab2cbf87ffb",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "df = df.loc[ (df['time']>dt.time(7)) & (df['time']<dt.time(16)) ]\n",
    "df = df[df.index.dayofweek < 5]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "33f5882d-46fb-4db4-85b7-6a36836ed1ad",
   "metadata": {},
   "source": [
    "The plot below shows the high daily variance and the gaps between data are non-trading hours"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 244,
   "id": "538e2e84-4524-420e-aea1-ee4f2c914e81",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Close_fut</th>\n",
       "      <th>Close_gas</th>\n",
       "      <th>Close_btc</th>\n",
       "      <th>Close_bond</th>\n",
       "      <th>date</th>\n",
       "      <th>time</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Local time</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2020-02-11 07:05:00</th>\n",
       "      <td>50.445</td>\n",
       "      <td>1.7640</td>\n",
       "      <td>9801</td>\n",
       "      <td>162.280</td>\n",
       "      <td>2020-02-11</td>\n",
       "      <td>07:05:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-02-11 07:10:00</th>\n",
       "      <td>50.375</td>\n",
       "      <td>1.7674</td>\n",
       "      <td>9815</td>\n",
       "      <td>162.311</td>\n",
       "      <td>2020-02-11</td>\n",
       "      <td>07:10:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-02-11 07:15:00</th>\n",
       "      <td>50.326</td>\n",
       "      <td>1.7703</td>\n",
       "      <td>9816</td>\n",
       "      <td>162.248</td>\n",
       "      <td>2020-02-11</td>\n",
       "      <td>07:15:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-02-11 07:20:00</th>\n",
       "      <td>50.356</td>\n",
       "      <td>1.7725</td>\n",
       "      <td>9821</td>\n",
       "      <td>162.561</td>\n",
       "      <td>2020-02-11</td>\n",
       "      <td>07:20:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-02-11 07:25:00</th>\n",
       "      <td>50.445</td>\n",
       "      <td>1.7734</td>\n",
       "      <td>9816</td>\n",
       "      <td>162.401</td>\n",
       "      <td>2020-02-11</td>\n",
       "      <td>07:25:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2023-02-10 15:35:00</th>\n",
       "      <td>79.799</td>\n",
       "      <td>2.5718</td>\n",
       "      <td>21531</td>\n",
       "      <td>126.899</td>\n",
       "      <td>2023-02-10</td>\n",
       "      <td>15:35:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2023-02-10 15:40:00</th>\n",
       "      <td>79.779</td>\n",
       "      <td>2.5721</td>\n",
       "      <td>21487</td>\n",
       "      <td>126.899</td>\n",
       "      <td>2023-02-10</td>\n",
       "      <td>15:40:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2023-02-10 15:45:00</th>\n",
       "      <td>79.764</td>\n",
       "      <td>2.5739</td>\n",
       "      <td>21552</td>\n",
       "      <td>126.901</td>\n",
       "      <td>2023-02-10</td>\n",
       "      <td>15:45:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2023-02-10 15:50:00</th>\n",
       "      <td>79.739</td>\n",
       "      <td>2.5721</td>\n",
       "      <td>21537</td>\n",
       "      <td>126.941</td>\n",
       "      <td>2023-02-10</td>\n",
       "      <td>15:50:00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2023-02-10 15:55:00</th>\n",
       "      <td>79.734</td>\n",
       "      <td>2.5761</td>\n",
       "      <td>21473</td>\n",
       "      <td>126.928</td>\n",
       "      <td>2023-02-10</td>\n",
       "      <td>15:55:00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>83888 rows × 6 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                     Close_fut  Close_gas  Close_btc  Close_bond        date  \\\n",
       "Local time                                                                     \n",
       "2020-02-11 07:05:00     50.445     1.7640       9801     162.280  2020-02-11   \n",
       "2020-02-11 07:10:00     50.375     1.7674       9815     162.311  2020-02-11   \n",
       "2020-02-11 07:15:00     50.326     1.7703       9816     162.248  2020-02-11   \n",
       "2020-02-11 07:20:00     50.356     1.7725       9821     162.561  2020-02-11   \n",
       "2020-02-11 07:25:00     50.445     1.7734       9816     162.401  2020-02-11   \n",
       "...                        ...        ...        ...         ...         ...   \n",
       "2023-02-10 15:35:00     79.799     2.5718      21531     126.899  2023-02-10   \n",
       "2023-02-10 15:40:00     79.779     2.5721      21487     126.899  2023-02-10   \n",
       "2023-02-10 15:45:00     79.764     2.5739      21552     126.901  2023-02-10   \n",
       "2023-02-10 15:50:00     79.739     2.5721      21537     126.941  2023-02-10   \n",
       "2023-02-10 15:55:00     79.734     2.5761      21473     126.928  2023-02-10   \n",
       "\n",
       "                         time  \n",
       "Local time                     \n",
       "2020-02-11 07:05:00  07:05:00  \n",
       "2020-02-11 07:10:00  07:10:00  \n",
       "2020-02-11 07:15:00  07:15:00  \n",
       "2020-02-11 07:20:00  07:20:00  \n",
       "2020-02-11 07:25:00  07:25:00  \n",
       "...                       ...  \n",
       "2023-02-10 15:35:00  15:35:00  \n",
       "2023-02-10 15:40:00  15:40:00  \n",
       "2023-02-10 15:45:00  15:45:00  \n",
       "2023-02-10 15:50:00  15:50:00  \n",
       "2023-02-10 15:55:00  15:55:00  \n",
       "\n",
       "[83888 rows x 6 columns]"
      ]
     },
     "execution_count": 264,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 245,
   "id": "cabaa440-8add-4e3e-b931-bed89cb7cba2",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x7f72a5be3940>"
      ]
     },
     "execution_count": 265,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1440x648 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.subplots(figsize=(20,9))\n",
    "plt.scatter(df.index, df['Close_fut'])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "aab721b6-33f9-4ee1-b214-7a50bc48b80a",
   "metadata": {},
   "source": [
    "# Test Train Splits\n",
    "\n",
    "## brainstorm:\n",
    "2 train test splitting techniques: day by day or chunk by chunk.\n",
    "\n",
    "- day by day: train p of the day test 1-p for each day, record accuracy for each day \n",
    "    - bad: small amount of data per train\n",
    "    - good: consistent times predicted\n",
    "- chunk by chunk: train n data points test m data points, record accuracy for each chunk\n",
    "    - bad: inconsistent time predicted; different days have different vol levels and predicting across days could be problematic\n",
    "    - good: more train data, could get different prediction horizons, no dependency on prediction time"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bf49793b-c8ae-4bb6-98d6-8844a46e4a3a",
   "metadata": {},
   "source": [
    "## Day by Day"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c65e5605-0686-4eb0-8084-8e7bc2c16ef2",
   "metadata": {},
   "source": [
    "We split the data into individual days.  We calculate the $n$ period rolling standard deviation as our outcome measure within each day.  Then, we can concat the data together"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 246,
   "id": "ee78ffb2-11ef-4ab1-abe3-9236b140e20f",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "days = list(set([d.date() for d in df.index]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 247,
   "id": "604f3797-3454-4309-a5cf-ecb433d48999",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'Close_bond', 'Close_btc', 'Close_fut', 'Close_gas'}"
      ]
     },
     "execution_count": 267,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "set(df.columns) - {'date', 'time'}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 271,
   "id": "f20c041d-c6bd-47ee-a588-883c184c1356",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "n=12\n",
    "day_dict = {}\n",
    "for day in days:\n",
    "        \n",
    "    day_dict[day] = df.loc[df['date'] == day]\n",
    "    del day_dict[day]['date']\n",
    "    \n",
    "    for feat in set(df.columns) - {'date', 'time'}:\n",
    "        day_dict[day][feat] = np.log( day_dict[day][feat] / day_dict[day][feat].shift(1) )#.shift(1) #last shift new\n",
    "    \n",
    "    day_dict[day][str(n)+'_oil_std'] = day_dict[day].rolling(n)['Close_fut'].std()\n",
    "    day_dict[day] = day_dict[day][n:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 272,
   "id": "cae91e4c-3414-4a1d-b82b-e4e95b8e627c",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x7f728de8bc18>"
      ]
     },
     "execution_count": 272,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1440x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "concat = pd.concat(day_dict.values())\n",
    "plt.subplots(figsize=(20,10))\n",
    "plt.scatter(range(700), concat[:700].reset_index(drop=True)[str(n)+'_oil_std'])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "59b7a930-e2b4-4b61-84a4-3da8fdd0b1c7",
   "metadata": {},
   "source": [
    "## Examining Temporal Structure"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 273,
   "id": "575df2ba-0cd8-4cd7-aac4-b732ea7374ca",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "time_dist = concat.groupby('time')[str(n)+'_oil_std'].mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 274,
   "id": "1be939c9-cd73-4c37-a9b8-a30f267018c3",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:xlabel='time'>"
      ]
     },
     "execution_count": 274,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "time_dist.plot()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "50efb558-d3f0-4704-8dc0-c038f7cb607e",
   "metadata": {},
   "source": [
    "The above plot shows that the trading volatility is clustered around the morning and roughly 2pm (historic oil desk close).  This is to be expected.  It will be important to strip times from the is dataset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 275,
   "id": "c2580609-4326-4f8b-95d1-090ec3d372c7",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "time_rets = concat.groupby('time')['Close_fut'].apply(lambda x: np.mean(abs(x)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 276,
   "id": "8cdfbfba-4790-43d3-89c0-01cac14f17dc",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:xlabel='time'>"
      ]
     },
     "execution_count": 276,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "time_rets.plot()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9d7a19e3-c9d0-4e62-ac5a-964d3af4b692",
   "metadata": {},
   "source": [
    "Again we see the largest log returns concentrated around 1:30 pm"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8adf7757-e946-4d5c-b0a8-0d6c2a8cdfeb",
   "metadata": {},
   "source": [
    "## Chunk by Chunk"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 277,
   "id": "99e4c9b9-2526-4e0b-98ef-cc42148ece30",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "784"
      ]
     },
     "execution_count": 277,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(days)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 278,
   "id": "28972497-3e01-4aba-aaf3-f1fbebb42eea",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "chunk_size = 1000"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 279,
   "id": "be3bb516-e712-4e4b-bec3-89b4d1e8c94b",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "chunk_dict = {}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 280,
   "id": "a4ce4146-196d-441b-84e9-1f99f2d1260c",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "for i,chunk in df.groupby(np.arange(len(df)) // chunk_size):\n",
    "    chunk_dict[i] = chunk.drop(['date', 'time'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 281,
   "id": "4e93b6da-aa30-43e6-a334-5e5b53647ade",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "chunk_indices = list(chunk_dict.keys())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 282,
   "id": "b9629e3c-aae8-4d24-9c55-6e5a0c9c2055",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "84"
      ]
     },
     "execution_count": 282,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(chunk_indices)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "23034976-b917-4e34-984f-c8c205802dbb",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "from scipy import stats\n",
    "from scipy.special import gamma\n",
    "from math import ceil, sqrt\n",
    "\n",
    "# Define the feature functions\n",
    "# Some constants for feature creation\n",
    "# Some functions adapted from https://github.com/BayerSe/RealizedQuantities\n",
    "mu_1 = np.sqrt((2 / np.pi))\n",
    "mu_43 = 2 ** (2 / 3) * gamma(7 / 6) * gamma(1 / 2) ** (-1)\n",
    "\n",
    "def rv(x):\n",
    "    return (x ** 2).sum()\n",
    "\n",
    "def rav(x):\n",
    "    return mu_1 ** (-1) * 1 ** (-.5) * x.abs().sum()\n",
    "\n",
    "def bv(x):\n",
    "    return mu_1 ** (-2) * (x.abs() * x.shift(1).abs()).sum()\n",
    "\n",
    "def tq(x):\n",
    "    return mu_43 ** (-3) *(x.abs() ** (4 / 3) * x.shift(1).abs() ** (4 / 3) * \n",
    "                                x.shift(2).abs() ** (4 / 3)).sum()\n",
    "    \n",
    "def jump_test(x):\n",
    "    j = ((np.log(rv(x)+0.00000001) - np.log(bv(x)+0.00000001)) / \\\n",
    "        ((mu_1 ** -4 + 2 * mu_1 ** -2 - 5) / (M * tq(x) * bv(x) ** -2)) ** 0.5)\n",
    "\n",
    "    jump = abs(j) >= stats.norm.ppf(0.999)\n",
    "    \n",
    "    return jump\n",
    "\n",
    "def rv_m(x):\n",
    "    return((x ** 2 * (x < 0)).sum())\n",
    "\n",
    "def rv_p(x):\n",
    "    return((x ** 2 * (x > 0)).sum())\n",
    "\n",
    "def sjv_p(x):\n",
    "    return((rv_p(x) - rv_m(x)) * (rv_p(x) - rv_m(x) > 0))\n",
    "\n",
    "def sjv_m(x):\n",
    "    return((rv_p(x) - rv_m(x)) * (rv_p(x) - rv_m(x) < 0))\n",
    "\n",
    "def rs(x):\n",
    "    return(np.sqrt(M) * (x ** 3).sum() / rv(x) ** (3 / 2))\n",
    "\n",
    "def rk(x):\n",
    "    return((x ** 4).sum() / rv(x) ** 2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 283,
   "id": "117aa8cf-37dd-4a6b-9b29-d3920b36f83d",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 19.2 s, sys: 38.8 ms, total: 19.2 s\n",
      "Wall time: 19.2 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "n=12\n",
    "\n",
    "for chunk in chunk_indices[:2]:    \n",
    "    for feat in set(df.columns) - {'date', 'time'}:\n",
    "        chunk_dict[chunk][feat] = np.log( chunk_dict[chunk][feat] / chunk_dict[chunk][feat].shift(1) ).shift(1)\n",
    "        \n",
    "    chunk_dict[chunk][str(n)+'_oil_std'] = chunk_dict[chunk].rolling(n)['Close_fut'].std()\n",
    "    \n",
    "    feat_funcs = [rv,rav,bv,tq,jump_test,rv_m,rv_p,sjv_p,sjv_m,rs,rk]\n",
    "\n",
    "    for func in feat_funcs:\n",
    "        # print(func)\n",
    "        chunk_dict[chunk][\"Fut_close_\"+func.__name__] = chunk_dict[chunk]['Close_fut'].rolling(n).apply(func).shift(1)\n",
    "    \n",
    "    chunk_dict[chunk] = chunk_dict[chunk][n:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 285,
   "id": "60c6a33b-dc41-4143-9132-5c4433863637",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x7f72a3cf5358>"
      ]
     },
     "execution_count": 285,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1440x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "concat = pd.concat(day_dict.values())\n",
    "plt.subplots(figsize=(20,10))\n",
    "plt.scatter(range(300), concat[:300].reset_index(drop=True)[str(n)+'_oil_std'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 286,
   "id": "3e3841eb-c231-48b0-8ced-58564f2bbaa3",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x7f7256e15e80>"
      ]
     },
     "execution_count": 286,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1440x720 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "concat = pd.concat(chunk_dict.values())\n",
    "plt.subplots(figsize=(20,10))\n",
    "plt.scatter(range(300), concat[:300].reset_index(drop=True)[str(n)+'_oil_std'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "1dae4c73-59a3-44bb-9e1f-5d067307e918",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-03-09T01:49:26.204757Z",
     "iopub.status.busy": "2023-03-09T01:49:26.204757Z",
     "iopub.status.idle": "2023-03-09T01:49:28.791678Z",
     "shell.execute_reply": "2023-03-09T01:49:28.790266Z",
     "shell.execute_reply.started": "2023-03-09T01:49:26.204757Z"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "concat.to_csv('test_data.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cdd94ad9-1748-43a0-8718-4a860e5803bd",
   "metadata": {},
   "source": [
    "We can see the difference in splitting methods.  New days will appear as large jumps like this rather than smooth as in the day splitting.  \n",
    "\n",
    "We might be able to fix this with some kind of smoothing technique ie windsorize and replace with some interpolation but it is probably best to just leave it be.  "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "feee45d2-2b38-4e5a-a475-216c62fde4e9",
   "metadata": {},
   "source": [
    "# Predicting"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "03570ea1-a801-4c5b-8dde-6c630befe39a",
   "metadata": {},
   "source": [
    "We train on a fraction $p$ of each chunk and test on the remaining $1-p$ of the chunk."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 288,
   "id": "2a624d77-c6fc-4074-adb0-15876aa9b84a",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "df=pd.read_csv('test_data.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 289,
   "id": "fad8f25d-5566-46c1-9444-7d494e7e52f5",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Local time</th>\n",
       "      <th>Close_fut</th>\n",
       "      <th>Close_gas</th>\n",
       "      <th>Close_btc</th>\n",
       "      <th>Close_bond</th>\n",
       "      <th>12_oil_std</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2020-02-11 08:10:00</td>\n",
       "      <td>0.001783</td>\n",
       "      <td>0.000057</td>\n",
       "      <td>0.000610</td>\n",
       "      <td>-0.000191</td>\n",
       "      <td>0.001318</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2020-02-11 08:15:00</td>\n",
       "      <td>0.001602</td>\n",
       "      <td>0.001810</td>\n",
       "      <td>0.000914</td>\n",
       "      <td>-0.000203</td>\n",
       "      <td>0.001308</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2020-02-11 08:20:00</td>\n",
       "      <td>-0.000415</td>\n",
       "      <td>-0.000170</td>\n",
       "      <td>-0.000203</td>\n",
       "      <td>0.000424</td>\n",
       "      <td>0.001332</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2020-02-11 08:25:00</td>\n",
       "      <td>0.000613</td>\n",
       "      <td>0.003836</td>\n",
       "      <td>-0.001016</td>\n",
       "      <td>0.000381</td>\n",
       "      <td>0.001264</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2020-02-11 08:30:00</td>\n",
       "      <td>-0.003978</td>\n",
       "      <td>0.000675</td>\n",
       "      <td>-0.001730</td>\n",
       "      <td>0.000350</td>\n",
       "      <td>0.001754</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>82791</th>\n",
       "      <td>2023-02-10 15:35:00</td>\n",
       "      <td>0.000251</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-0.000696</td>\n",
       "      <td>-0.000252</td>\n",
       "      <td>0.000467</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>82792</th>\n",
       "      <td>2023-02-10 15:40:00</td>\n",
       "      <td>-0.000251</td>\n",
       "      <td>0.000117</td>\n",
       "      <td>-0.002046</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000418</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>82793</th>\n",
       "      <td>2023-02-10 15:45:00</td>\n",
       "      <td>-0.000188</td>\n",
       "      <td>0.000700</td>\n",
       "      <td>0.003021</td>\n",
       "      <td>0.000016</td>\n",
       "      <td>0.000419</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>82794</th>\n",
       "      <td>2023-02-10 15:50:00</td>\n",
       "      <td>-0.000313</td>\n",
       "      <td>-0.000700</td>\n",
       "      <td>-0.000696</td>\n",
       "      <td>0.000315</td>\n",
       "      <td>0.000422</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>82795</th>\n",
       "      <td>2023-02-10 15:55:00</td>\n",
       "      <td>-0.000063</td>\n",
       "      <td>0.001554</td>\n",
       "      <td>-0.002976</td>\n",
       "      <td>-0.000102</td>\n",
       "      <td>0.000365</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>82796 rows × 6 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                Local time  Close_fut  Close_gas  Close_btc  Close_bond  \\\n",
       "0      2020-02-11 08:10:00   0.001783   0.000057   0.000610   -0.000191   \n",
       "1      2020-02-11 08:15:00   0.001602   0.001810   0.000914   -0.000203   \n",
       "2      2020-02-11 08:20:00  -0.000415  -0.000170  -0.000203    0.000424   \n",
       "3      2020-02-11 08:25:00   0.000613   0.003836  -0.001016    0.000381   \n",
       "4      2020-02-11 08:30:00  -0.003978   0.000675  -0.001730    0.000350   \n",
       "...                    ...        ...        ...        ...         ...   \n",
       "82791  2023-02-10 15:35:00   0.000251   0.000000  -0.000696   -0.000252   \n",
       "82792  2023-02-10 15:40:00  -0.000251   0.000117  -0.002046    0.000000   \n",
       "82793  2023-02-10 15:45:00  -0.000188   0.000700   0.003021    0.000016   \n",
       "82794  2023-02-10 15:50:00  -0.000313  -0.000700  -0.000696    0.000315   \n",
       "82795  2023-02-10 15:55:00  -0.000063   0.001554  -0.002976   -0.000102   \n",
       "\n",
       "       12_oil_std  \n",
       "0        0.001318  \n",
       "1        0.001308  \n",
       "2        0.001332  \n",
       "3        0.001264  \n",
       "4        0.001754  \n",
       "...           ...  \n",
       "82791    0.000467  \n",
       "82792    0.000418  \n",
       "82793    0.000419  \n",
       "82794    0.000422  \n",
       "82795    0.000365  \n",
       "\n",
       "[82796 rows x 6 columns]"
      ]
     },
     "execution_count": 289,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "961f5572-7de0-4ab8-8117-6f11cafe0f6c",
   "metadata": {},
   "source": [
    "We get a train test split for each chunk."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "id": "0d4089b3-19f2-45b9-9fd0-342534b7ffe6",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "chunk_size = 1000\n",
    "data_chunks = {}\n",
    "\n",
    "for i,chunk in df.groupby(np.arange(len(df)) // chunk_size):\n",
    "    total = chunk#.drop(['date', 'time'], axis=1)\n",
    "    \n",
    "    X = total[['Close_gas', 'Close_btc', 'Close_bond', 'Close_fut']]\n",
    "    Y = total['12_oil_std']\n",
    "    \n",
    "    X_train, X_test, y_train, y_test = train_test_split(X, Y, test_size=0.1, shuffle=False)\n",
    "    \n",
    "    data_chunks[i] = [X_train, X_test, y_train, y_test]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "id": "754d0c04-4a48-43a6-b038-31a8eb027d9e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "83"
      ]
     },
     "execution_count": 81,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(data_chunks)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3b31a141-ee4c-445e-80ef-d1930cccc35e",
   "metadata": {},
   "source": [
    "## Predicting by chunk"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "id": "0f1a30a1-30f8-42e0-a9b9-bb2c6de6219c",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "from sklearn.model_selection import KFold, StratifiedKFold, train_test_split\n",
    "from sklearn.linear_model import ElasticNet, Ridge, Lasso\n",
    "from sklearn.ensemble import RandomForestRegressor, GradientBoostingRegressor, AdaBoostRegressor\n",
    "from sklearn.kernel_ridge import KernelRidge\n",
    "from sklearn.neighbors import KNeighborsRegressor\n",
    "from sklearn.metrics import mean_absolute_error, accuracy_score, mean_squared_error\n",
    "from sklearn.model_selection import TimeSeriesSplit, GridSearchCV, cross_val_score, KFold, StratifiedKFold\n",
    "from sklearn_extensions.kernel_regression import KernelRegression\n",
    "from statsmodels.graphics.tsaplots import plot_acf, plot_pacf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "dbf9f61d-0b29-4977-a632-665deaf0102d",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "# A bunch of nonparametric regression models + some parametric baselines\n",
    "def model_selection(ml, data_list):\n",
    "    X_train, X_test, y_train, y_test = data_list[0], data_list[1], data_list[2], data_list[3]\n",
    "    # print(ml)\n",
    "    \n",
    "    cores=35\n",
    "    \n",
    "    model_best = ''\n",
    "    if ml == 'GB':\n",
    "        model = GradientBoostingRegressor()\n",
    "        grid = dict()\n",
    "        grid['learning_rate'] = [1e-3, 2*1e-3, 3*1e-3, 4*1e-3, 5*1e-3, 7*1e-3] #\n",
    "        grid['n_estimators'] = [1000,2000] \n",
    "        grid_search = GridSearchCV(estimator=model, param_grid=grid, scoring='neg_mean_squared_error', n_jobs=cores)\n",
    "        grid_result = grid_search.fit(X_train, y_train)\n",
    "        model_best = GradientBoostingRegressor(learning_rate=grid_result.best_params_['learning_rate'],\\\n",
    "                    n_estimators=grid_result.best_params_['n_estimators'],\\\n",
    "                                              ).fit(X_train, y_train)\n",
    "\n",
    "    elif ml == 'RF':\n",
    "        model = RandomForestRegressor(n_jobs=cores)\n",
    "        grid = dict()\n",
    "        grid['max_depth'] = [3,4,5,7]\n",
    "        grid['n_estimators'] = [1000,2000]\n",
    "        grid_search = GridSearchCV(estimator=model, param_grid=grid, scoring='neg_mean_squared_error', n_jobs=cores)\n",
    "        grid_result = grid_search.fit(X_train, y_train)\n",
    "\n",
    "        model_best = RandomForestRegressor(\n",
    "                    max_depth=grid_result.best_params_['max_depth'],\\\n",
    "                    n_estimators=grid_result.best_params_['n_estimators'],\\\n",
    "\n",
    "                        ).fit(X_train, y_train)\n",
    "    elif ml == 'AB':\n",
    "        model = AdaBoostRegressor()\n",
    "        grid = dict()\n",
    "        grid['learning_rate'] = [1e-4,1e-2,1]\n",
    "        grid['n_estimators'] = [100,1000]\n",
    "        grid_search = GridSearchCV(estimator=model, param_grid=grid, scoring='neg_mean_squared_error', n_jobs=cores)\n",
    "        grid_result = grid_search.fit(X_train, y_train)\n",
    "\n",
    "        model_best = AdaBoostRegressor(\n",
    "                    learning_rate=grid_result.best_params_['learning_rate'],\\\n",
    "                    n_estimators=grid_result.best_params_['n_estimators'],\\\n",
    "                        ).fit(X_train, y_train)\n",
    "          \n",
    "    elif ml == 'KR':\n",
    "        model = KernelRidge()\n",
    "        grid = dict()\n",
    "        grid['alpha'] = [1e-3,1e-2,1e-1]\n",
    "        grid[\"kernel\"] = [\"poly\", \"cosine\"]\n",
    "        grid_search = GridSearchCV(estimator=model, param_grid=grid, scoring='neg_mean_squared_error', n_jobs=cores)\n",
    "        grid_result = grid_search.fit(X_train, y_train)\n",
    "        model_best = KernelRidge(alpha=grid_result.best_params_[\"alpha\"],\n",
    "                                 kernel=grid_result.best_params_[\"kernel\"]).fit(X_train, y_train)\n",
    "    elif ml == 'R':\n",
    "        model = Ridge()\n",
    "        grid = dict()\n",
    "        grid['alpha'] = [1e-2,1e-3,1e-1]\n",
    "        grid_search = GridSearchCV(estimator=model, param_grid=grid, scoring='neg_mean_squared_error', n_jobs=cores)\n",
    "        grid_result = grid_search.fit(X_train, y_train)\n",
    "        model_best = Ridge(alpha=grid_result.best_params_[\"alpha\"]).fit(X_train, y_train)\n",
    "        \n",
    "    elif ml == \"LA\":\n",
    "        model = Lasso()\n",
    "        grid = dict()\n",
    "        grid['alpha'] = [1e-6,1e-5, 1e-4, 1e-3]\n",
    "        grid_search = GridSearchCV(estimator=model, param_grid=grid, scoring='neg_mean_squared_error', n_jobs=cores)\n",
    "        grid_result = grid_search.fit(X_train, y_train)\n",
    "        model_best = Lasso(alpha=grid_result.best_params_[\"alpha\"]).fit(X_train, y_train)\n",
    "        \n",
    "        \n",
    "    elif ml == \"EN\":\n",
    "        model = ElasticNet()\n",
    "        grid = dict()\n",
    "        grid['alpha'] = [1e-6,1e-5, 1e-4, 1e-3]\n",
    "        grid[\"l1_ratio\"] = [0.1,0.5,0.7,0.9,0.95]\n",
    "        grid_search = GridSearchCV(estimator=model, param_grid=grid, scoring='neg_mean_squared_error', n_jobs=cores)\n",
    "        grid_result = grid_search.fit(X_train, y_train)\n",
    "        model_best = ElasticNet(alpha=grid_result.best_params_[\"alpha\"],\n",
    "                                l1_ratio = grid_result.best_params_[\"l1_ratio\"]).fit(X_train, y_train)\n",
    "    \n",
    "    elif ml == \"NW\":\n",
    "        # Bandwidth is selected automatically\n",
    "        model = KernelRegression()\n",
    "        grid = dict()\n",
    "        grid[\"gamma\"] = [1e-2, 1e-1, 1, 10,100]\n",
    "        grid['kernel'] = [\"rbf\",\"poly\"]\n",
    "        grid_search = GridSearchCV(estimator=model, param_grid=grid, scoring='neg_mean_squared_error', n_jobs=cores)\n",
    "        grid_result = grid_search.fit(X_train, y_train)\n",
    "        model_best = KernelRegression(gamma=grid_result.best_params_[\"gamma\"],\n",
    "                            kernel = grid_result.best_params_[\"kernel\"]).fit(X_train, y_train)\n",
    "        \n",
    "    elif ml == \"KN\":\n",
    "        model = KNeighborsRegressor()\n",
    "        grid = dict()\n",
    "        grid[\"n_neighbors\"] = [5,10,40,100]\n",
    "        grid['weights'] = [\"uniform\",\"distance\"]\n",
    "        grid_search = GridSearchCV(estimator=model, param_grid=grid, scoring='neg_mean_squared_error', n_jobs=cores)\n",
    "        grid_result = grid_search.fit(X_train, y_train)\n",
    "        model_best = KNeighborsRegressor(n_neighbors=grid_result.best_params_[\"n_neighbors\"],\n",
    "                            weights = grid_result.best_params_[\"weights\"]).fit(X_train, y_train)\n",
    "        \n",
    "    # print(model_best)\n",
    "    y_pred = model_best.predict(X_test)\n",
    "    acc = {'MAE':mean_absolute_error(y_pred, y_test), 'MAPE':100 * mean_absolute_error(y_pred, y_test)/np.mean(y_test), \n",
    "           'MSE':mean_squared_error(y_pred,y_test), 'RMSE':np.sqrt(mean_squared_error(y_pred,y_test))}\n",
    "    \n",
    "    # print('MAE: ', mean_absolute_error(y_pred, y_test), \n",
    "    #       'MAE%: ', 100 * mean_absolute_error(y_pred, y_test)/np.mean(y_test), \n",
    "    #       '\\nMSE: ', mean_squared_error(y_pred,y_test), \n",
    "    #       '\\nRMSE: ', np.sqrt(mean_squared_error(y_pred,y_test)))\n",
    "\n",
    "    return(model_best, y_test, y_pred, acc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "id": "6fdef8a5-a2ae-4274-a7ef-30125e4e26bc",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0 0:00:21.928838\n",
      "1 0:00:18.526302\n",
      "2 0:00:21.332554\n",
      "3 0:00:18.624204\n",
      "4 0:00:26.157691\n",
      "5 0:00:15.699685\n",
      "6 0:00:18.215578\n",
      "7 0:00:22.903170\n",
      "8 0:00:20.562592\n",
      "9 0:00:21.134479\n",
      "10 0:00:20.925249\n",
      "11 0:00:17.225143\n",
      "12 0:00:15.430487\n",
      "13 0:00:15.884101\n",
      "14 0:00:21.191981\n",
      "15 0:00:18.533857\n",
      "16 0:00:18.116262\n",
      "17 0:00:15.873829\n",
      "18 0:00:18.627627\n",
      "19 0:00:23.043267\n",
      "20 0:00:21.415202\n",
      "21 0:00:18.122051\n",
      "22 0:00:15.387737\n",
      "23 0:00:16.003321\n",
      "24 0:00:19.022791\n",
      "25 0:00:24.466780\n",
      "26 0:00:17.158539\n",
      "27 0:00:20.384312\n",
      "28 0:00:20.117606\n",
      "29 0:00:20.735002\n",
      "30 0:00:22.192032\n",
      "31 0:00:17.716973\n",
      "32 0:00:23.438790\n",
      "33 0:00:20.603035\n",
      "34 0:00:15.481415\n",
      "35 0:00:15.590423\n",
      "36 0:00:19.893621\n",
      "37 0:00:18.939029\n",
      "38 0:00:17.861055\n",
      "39 0:00:15.728581\n",
      "40 0:00:24.404964\n",
      "41 0:00:23.457765\n",
      "42 0:00:23.219970\n",
      "43 0:00:21.157231\n",
      "44 0:00:14.119720\n",
      "45 0:00:21.480059\n",
      "46 0:00:17.840185\n",
      "47 0:00:20.758689\n",
      "48 0:00:22.369471\n",
      "49 0:00:20.592920\n",
      "50 0:00:23.653909\n",
      "51 0:00:18.484810\n",
      "52 0:00:23.625178\n",
      "53 0:00:23.849089\n",
      "54 0:00:18.729793\n",
      "55 0:00:15.903829\n",
      "56 0:00:15.961779\n",
      "57 0:00:19.682729\n",
      "58 0:00:15.468394\n",
      "59 0:00:23.181968\n",
      "60 0:00:18.227808\n",
      "61 0:00:13.884833\n",
      "62 0:00:20.038216\n",
      "63 0:00:15.478073\n",
      "64 0:00:17.779613\n",
      "65 0:00:18.183932\n",
      "66 0:00:23.213397\n",
      "67 0:00:15.571424\n",
      "68 0:00:17.570870\n",
      "69 0:00:22.772058\n",
      "70 0:00:17.644136\n",
      "71 0:00:15.867418\n",
      "72 0:00:20.275182\n",
      "73 0:00:15.688961\n",
      "74 0:00:20.675149\n",
      "75 0:00:20.194720\n",
      "76 0:00:15.625534\n",
      "77 0:00:20.610935\n",
      "78 0:00:17.579758\n",
      "79 0:00:19.231680\n",
      "80 0:00:18.712801\n",
      "81 0:00:20.823648\n",
      "82 0:00:20.502802\n",
      "CPU times: user 11min 24s, sys: 25 s, total: 11min 49s\n",
      "Wall time: 26min 44s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "chunk_results = {}\n",
    "indices = data_chunks.keys()\n",
    "\n",
    "indices = list(indices)\n",
    "\n",
    "mod_list = ['GB', 'RF', 'AB', 'KR', 'R', 'LA', 'EN', 'NW', 'KN']\n",
    "\n",
    "for i in indices:\n",
    "    t = dt.datetime.now()\n",
    "    chunk_results[i] = {}\n",
    "    for ml in mod_list:\n",
    "        chunk_results[i][ml] = model_selection(ml, data_chunks[i])\n",
    "    print(i, dt.datetime.now()-t)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "id": "0885e37a-8897-4fa3-b773-b43e2e6994eb",
   "metadata": {},
   "outputs": [],
   "source": [
    "for label in chunk_results.keys():\n",
    "    chunk_results[label]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "id": "6d1cbcad-1f2d-494d-a854-74dd8728d879",
   "metadata": {},
   "outputs": [],
   "source": [
    "acc_dict = {}\n",
    "for chunk in chunk_results.keys():\n",
    "    for m in mod_list:\n",
    "        acc_dict[str(chunk)+m] = chunk_results[chunk][m][-1]\n",
    "        acc_dict[str(chunk)+m]['model'] = m\n",
    "        # acc_dict[m].append(m)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "id": "a26eca7d-7742-416d-b5b1-a34ce93c7325",
   "metadata": {},
   "outputs": [],
   "source": [
    "acc_df = pd.DataFrame(acc_dict).transpose()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "id": "9811ac37-b553-4872-95b6-ff67861f3363",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>MAE</th>\n",
       "      <th>MAPE</th>\n",
       "      <th>MSE</th>\n",
       "      <th>RMSE</th>\n",
       "      <th>model</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0GB</th>\n",
       "      <td>0.00155702</td>\n",
       "      <td>68.1247</td>\n",
       "      <td>1.06173e-05</td>\n",
       "      <td>0.00325843</td>\n",
       "      <td>GB</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0RF</th>\n",
       "      <td>0.00159527</td>\n",
       "      <td>69.798</td>\n",
       "      <td>1.09087e-05</td>\n",
       "      <td>0.00330283</td>\n",
       "      <td>RF</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0AB</th>\n",
       "      <td>0.00158904</td>\n",
       "      <td>69.5257</td>\n",
       "      <td>1.13249e-05</td>\n",
       "      <td>0.00336525</td>\n",
       "      <td>AB</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0KR</th>\n",
       "      <td>0.00162457</td>\n",
       "      <td>71.08</td>\n",
       "      <td>1.20572e-05</td>\n",
       "      <td>0.00347235</td>\n",
       "      <td>KR</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0R</th>\n",
       "      <td>0.00162563</td>\n",
       "      <td>71.1267</td>\n",
       "      <td>1.20537e-05</td>\n",
       "      <td>0.00347184</td>\n",
       "      <td>R</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>82R</th>\n",
       "      <td>0.000782065</td>\n",
       "      <td>72.4023</td>\n",
       "      <td>7.98161e-07</td>\n",
       "      <td>0.000893398</td>\n",
       "      <td>R</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>82LA</th>\n",
       "      <td>0.000782065</td>\n",
       "      <td>72.4023</td>\n",
       "      <td>7.98161e-07</td>\n",
       "      <td>0.000893399</td>\n",
       "      <td>LA</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>82EN</th>\n",
       "      <td>0.000782065</td>\n",
       "      <td>72.4023</td>\n",
       "      <td>7.98161e-07</td>\n",
       "      <td>0.000893399</td>\n",
       "      <td>EN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>82NW</th>\n",
       "      <td>0.000781433</td>\n",
       "      <td>72.3437</td>\n",
       "      <td>7.96861e-07</td>\n",
       "      <td>0.000892671</td>\n",
       "      <td>NW</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>82KN</th>\n",
       "      <td>0.000561127</td>\n",
       "      <td>51.9482</td>\n",
       "      <td>4.56408e-07</td>\n",
       "      <td>0.000675579</td>\n",
       "      <td>KN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>747 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "              MAE     MAPE          MSE         RMSE model\n",
       "0GB    0.00155702  68.1247  1.06173e-05   0.00325843    GB\n",
       "0RF    0.00159527   69.798  1.09087e-05   0.00330283    RF\n",
       "0AB    0.00158904  69.5257  1.13249e-05   0.00336525    AB\n",
       "0KR    0.00162457    71.08  1.20572e-05   0.00347235    KR\n",
       "0R     0.00162563  71.1267  1.20537e-05   0.00347184     R\n",
       "...           ...      ...          ...          ...   ...\n",
       "82R   0.000782065  72.4023  7.98161e-07  0.000893398     R\n",
       "82LA  0.000782065  72.4023  7.98161e-07  0.000893399    LA\n",
       "82EN  0.000782065  72.4023  7.98161e-07  0.000893399    EN\n",
       "82NW  0.000781433  72.3437  7.96861e-07  0.000892671    NW\n",
       "82KN  0.000561127  51.9482  4.56408e-07  0.000675579    KN\n",
       "\n",
       "[747 rows x 5 columns]"
      ]
     },
     "execution_count": 113,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "acc_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "id": "74c3adfc-301a-43f2-8425-a74316731126",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "model\n",
       "AB    48.891872\n",
       "EN    53.630768\n",
       "GB    49.292161\n",
       "KN    46.353923\n",
       "KR    53.783620\n",
       "LA    53.630768\n",
       "NW    53.408520\n",
       "R     53.755003\n",
       "RF    48.823544\n",
       "Name: MAPE, dtype: float64"
      ]
     },
     "execution_count": 112,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "acc_df.reset_index(drop=True).groupby('model')['MAPE'].apply(lambda x: np.mean(x))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9e930d25-7b56-431f-a2ad-94e9436ec507",
   "metadata": {},
   "source": [
    "# WIth new features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 290,
   "id": "0cdebc9c-ce88-482e-86d1-add69c6bd7ac",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "df=pd.read_csv('combined_chunks.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 291,
   "id": "fa875df2-c07c-4e65-a0df-88c5028e38c3",
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df.fillna(method='ffill')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 293,
   "id": "505a1879-e3a4-42f2-9a51-c0c7502bc607",
   "metadata": {},
   "outputs": [],
   "source": [
    "for feat in ['Close_fut', 'Close_gas', 'Close_btc', 'Close_bond']:\n",
    "    df[feat] = df[feat].shift(1)\n",
    "df = df[1:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 294,
   "id": "4def610d-b169-4dd7-a1f9-15362ce0a667",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Local time</th>\n",
       "      <th>Close_fut</th>\n",
       "      <th>Close_gas</th>\n",
       "      <th>Close_btc</th>\n",
       "      <th>Close_bond</th>\n",
       "      <th>12_oil_std</th>\n",
       "      <th>Fut_close_rv</th>\n",
       "      <th>Fut_close_rav</th>\n",
       "      <th>Fut_close_bv</th>\n",
       "      <th>Fut_close_tq</th>\n",
       "      <th>Fut_close_jump_test</th>\n",
       "      <th>Fut_close_rv_m</th>\n",
       "      <th>Fut_close_rv_p</th>\n",
       "      <th>Fut_close_sjv_p</th>\n",
       "      <th>Fut_close_sjv_m</th>\n",
       "      <th>Fut_close_rs</th>\n",
       "      <th>Fut_close_rk</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2020-02-11 08:10:00</td>\n",
       "      <td>-0.001189</td>\n",
       "      <td>-0.004012</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000775</td>\n",
       "      <td>0.001318</td>\n",
       "      <td>0.000017</td>\n",
       "      <td>0.014619</td>\n",
       "      <td>1.601874e-05</td>\n",
       "      <td>2.179593e-11</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000006</td>\n",
       "      <td>1.122443e-05</td>\n",
       "      <td>0.000006</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.228043</td>\n",
       "      <td>0.188860</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2020-02-11 08:15:00</td>\n",
       "      <td>0.001783</td>\n",
       "      <td>0.000057</td>\n",
       "      <td>0.000610</td>\n",
       "      <td>-0.000191</td>\n",
       "      <td>0.001308</td>\n",
       "      <td>0.000019</td>\n",
       "      <td>0.015634</td>\n",
       "      <td>1.843612e-05</td>\n",
       "      <td>1.999619e-11</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000005</td>\n",
       "      <td>1.440185e-05</td>\n",
       "      <td>0.000010</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.268589</td>\n",
       "      <td>0.172602</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2020-02-11 08:20:00</td>\n",
       "      <td>0.001602</td>\n",
       "      <td>0.001810</td>\n",
       "      <td>0.000914</td>\n",
       "      <td>-0.000203</td>\n",
       "      <td>0.001332</td>\n",
       "      <td>0.000021</td>\n",
       "      <td>0.016894</td>\n",
       "      <td>2.126749e-05</td>\n",
       "      <td>2.751696e-11</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000005</td>\n",
       "      <td>1.661173e-05</td>\n",
       "      <td>0.000012</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.267611</td>\n",
       "      <td>0.152855</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2020-02-11 08:25:00</td>\n",
       "      <td>-0.000415</td>\n",
       "      <td>-0.000170</td>\n",
       "      <td>-0.000203</td>\n",
       "      <td>0.000424</td>\n",
       "      <td>0.001264</td>\n",
       "      <td>0.000018</td>\n",
       "      <td>0.015201</td>\n",
       "      <td>2.011112e-05</td>\n",
       "      <td>2.546465e-11</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000005</td>\n",
       "      <td>1.349348e-05</td>\n",
       "      <td>0.000009</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.263453</td>\n",
       "      <td>0.177083</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>2020-02-11 08:30:00</td>\n",
       "      <td>0.000613</td>\n",
       "      <td>0.003836</td>\n",
       "      <td>-0.001016</td>\n",
       "      <td>0.000381</td>\n",
       "      <td>0.001754</td>\n",
       "      <td>0.000018</td>\n",
       "      <td>0.014974</td>\n",
       "      <td>1.877874e-05</td>\n",
       "      <td>1.967731e-11</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000004</td>\n",
       "      <td>1.386866e-05</td>\n",
       "      <td>0.000010</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.278580</td>\n",
       "      <td>0.181327</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>82875</th>\n",
       "      <td>2023-02-10 15:35:00</td>\n",
       "      <td>-0.000376</td>\n",
       "      <td>0.000856</td>\n",
       "      <td>-0.002596</td>\n",
       "      <td>0.000008</td>\n",
       "      <td>0.000467</td>\n",
       "      <td>0.000002</td>\n",
       "      <td>0.005073</td>\n",
       "      <td>1.930419e-06</td>\n",
       "      <td>1.946174e-13</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000002</td>\n",
       "      <td>7.263531e-07</td>\n",
       "      <td>-0.000000</td>\n",
       "      <td>-9.151377e-07</td>\n",
       "      <td>-0.240425</td>\n",
       "      <td>0.225858</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>82876</th>\n",
       "      <td>2023-02-10 15:40:00</td>\n",
       "      <td>0.000251</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>-0.000696</td>\n",
       "      <td>-0.000252</td>\n",
       "      <td>0.000418</td>\n",
       "      <td>0.000002</td>\n",
       "      <td>0.004445</td>\n",
       "      <td>1.930485e-06</td>\n",
       "      <td>2.125854e-13</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000001</td>\n",
       "      <td>7.891841e-07</td>\n",
       "      <td>-0.000000</td>\n",
       "      <td>-2.872951e-07</td>\n",
       "      <td>-0.170921</td>\n",
       "      <td>0.273229</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>82877</th>\n",
       "      <td>2023-02-10 15:45:00</td>\n",
       "      <td>-0.000251</td>\n",
       "      <td>0.000117</td>\n",
       "      <td>-0.002046</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000419</td>\n",
       "      <td>0.000002</td>\n",
       "      <td>0.004602</td>\n",
       "      <td>1.979839e-06</td>\n",
       "      <td>2.151393e-13</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000001</td>\n",
       "      <td>7.734795e-07</td>\n",
       "      <td>-0.000000</td>\n",
       "      <td>-3.658307e-07</td>\n",
       "      <td>-0.171341</td>\n",
       "      <td>0.260943</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>82878</th>\n",
       "      <td>2023-02-10 15:50:00</td>\n",
       "      <td>-0.000188</td>\n",
       "      <td>0.000700</td>\n",
       "      <td>0.003021</td>\n",
       "      <td>0.000016</td>\n",
       "      <td>0.000422</td>\n",
       "      <td>0.000002</td>\n",
       "      <td>0.004524</td>\n",
       "      <td>1.807205e-06</td>\n",
       "      <td>8.864888e-14</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000001</td>\n",
       "      <td>7.734795e-07</td>\n",
       "      <td>-0.000000</td>\n",
       "      <td>-3.383624e-07</td>\n",
       "      <td>-0.171584</td>\n",
       "      <td>0.267843</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>82879</th>\n",
       "      <td>2023-02-10 15:55:00</td>\n",
       "      <td>-0.000313</td>\n",
       "      <td>-0.000700</td>\n",
       "      <td>-0.000696</td>\n",
       "      <td>0.000315</td>\n",
       "      <td>0.000365</td>\n",
       "      <td>0.000002</td>\n",
       "      <td>0.004132</td>\n",
       "      <td>9.994822e-07</td>\n",
       "      <td>9.496995e-14</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000001</td>\n",
       "      <td>3.809629e-07</td>\n",
       "      <td>-0.000000</td>\n",
       "      <td>-8.291448e-07</td>\n",
       "      <td>-0.359202</td>\n",
       "      <td>0.319027</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>82879 rows × 17 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                Local time  Close_fut  Close_gas  Close_btc  Close_bond  \\\n",
       "1      2020-02-11 08:10:00  -0.001189  -0.004012   0.000000    0.000775   \n",
       "2      2020-02-11 08:15:00   0.001783   0.000057   0.000610   -0.000191   \n",
       "3      2020-02-11 08:20:00   0.001602   0.001810   0.000914   -0.000203   \n",
       "4      2020-02-11 08:25:00  -0.000415  -0.000170  -0.000203    0.000424   \n",
       "5      2020-02-11 08:30:00   0.000613   0.003836  -0.001016    0.000381   \n",
       "...                    ...        ...        ...        ...         ...   \n",
       "82875  2023-02-10 15:35:00  -0.000376   0.000856  -0.002596    0.000008   \n",
       "82876  2023-02-10 15:40:00   0.000251   0.000000  -0.000696   -0.000252   \n",
       "82877  2023-02-10 15:45:00  -0.000251   0.000117  -0.002046    0.000000   \n",
       "82878  2023-02-10 15:50:00  -0.000188   0.000700   0.003021    0.000016   \n",
       "82879  2023-02-10 15:55:00  -0.000313  -0.000700  -0.000696    0.000315   \n",
       "\n",
       "       12_oil_std  Fut_close_rv  Fut_close_rav  Fut_close_bv  Fut_close_tq  \\\n",
       "1        0.001318      0.000017       0.014619  1.601874e-05  2.179593e-11   \n",
       "2        0.001308      0.000019       0.015634  1.843612e-05  1.999619e-11   \n",
       "3        0.001332      0.000021       0.016894  2.126749e-05  2.751696e-11   \n",
       "4        0.001264      0.000018       0.015201  2.011112e-05  2.546465e-11   \n",
       "5        0.001754      0.000018       0.014974  1.877874e-05  1.967731e-11   \n",
       "...           ...           ...            ...           ...           ...   \n",
       "82875    0.000467      0.000002       0.005073  1.930419e-06  1.946174e-13   \n",
       "82876    0.000418      0.000002       0.004445  1.930485e-06  2.125854e-13   \n",
       "82877    0.000419      0.000002       0.004602  1.979839e-06  2.151393e-13   \n",
       "82878    0.000422      0.000002       0.004524  1.807205e-06  8.864888e-14   \n",
       "82879    0.000365      0.000002       0.004132  9.994822e-07  9.496995e-14   \n",
       "\n",
       "       Fut_close_jump_test  Fut_close_rv_m  Fut_close_rv_p  Fut_close_sjv_p  \\\n",
       "1                      0.0        0.000006    1.122443e-05         0.000006   \n",
       "2                      0.0        0.000005    1.440185e-05         0.000010   \n",
       "3                      0.0        0.000005    1.661173e-05         0.000012   \n",
       "4                      0.0        0.000005    1.349348e-05         0.000009   \n",
       "5                      0.0        0.000004    1.386866e-05         0.000010   \n",
       "...                    ...             ...             ...              ...   \n",
       "82875                  0.0        0.000002    7.263531e-07        -0.000000   \n",
       "82876                  0.0        0.000001    7.891841e-07        -0.000000   \n",
       "82877                  0.0        0.000001    7.734795e-07        -0.000000   \n",
       "82878                  0.0        0.000001    7.734795e-07        -0.000000   \n",
       "82879                  0.0        0.000001    3.809629e-07        -0.000000   \n",
       "\n",
       "       Fut_close_sjv_m  Fut_close_rs  Fut_close_rk  \n",
       "1         0.000000e+00      0.228043      0.188860  \n",
       "2         0.000000e+00      0.268589      0.172602  \n",
       "3         0.000000e+00      0.267611      0.152855  \n",
       "4         0.000000e+00      0.263453      0.177083  \n",
       "5         0.000000e+00      0.278580      0.181327  \n",
       "...                ...           ...           ...  \n",
       "82875    -9.151377e-07     -0.240425      0.225858  \n",
       "82876    -2.872951e-07     -0.170921      0.273229  \n",
       "82877    -3.658307e-07     -0.171341      0.260943  \n",
       "82878    -3.383624e-07     -0.171584      0.267843  \n",
       "82879    -8.291448e-07     -0.359202      0.319027  \n",
       "\n",
       "[82879 rows x 17 columns]"
      ]
     },
     "execution_count": 294,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 295,
   "id": "b940deeb-3d0e-41ac-9e6f-feba4278532c",
   "metadata": {},
   "outputs": [],
   "source": [
    "feats = list(set(df.columns) - set(['Local time', '12_oil_std', 'Fut_close_jump_test']))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 296,
   "id": "08947a61-010e-4201-b432-a046cdb18bb6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['Fut_close_bv',\n",
       " 'Fut_close_rk',\n",
       " 'Close_bond',\n",
       " 'Close_gas',\n",
       " 'Fut_close_rv_p',\n",
       " 'Fut_close_rv_m',\n",
       " 'Fut_close_sjv_p',\n",
       " 'Fut_close_rs',\n",
       " 'Close_btc',\n",
       " 'Fut_close_rv',\n",
       " 'Fut_close_sjv_m',\n",
       " 'Fut_close_tq',\n",
       " 'Close_fut',\n",
       " 'Fut_close_rav']"
      ]
     },
     "execution_count": 296,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "feats"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e17c3567-a260-416a-811d-d4949700f9b1",
   "metadata": {},
   "source": [
    "We get a train test split for each chunk."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 297,
   "id": "f818e06e-545e-40c2-8402-718cee7d4f61",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "chunk_size = 1000\n",
    "data_chunks = {}\n",
    "\n",
    "for i,chunk in df.groupby(np.arange(len(df)) // chunk_size):\n",
    "    total = chunk#.drop(['date', 'time'], axis=1)\n",
    "    \n",
    "    X = total[feats]\n",
    "    Y = total['12_oil_std']\n",
    "    \n",
    "    X_train, X_test, y_train, y_test = train_test_split(X, Y, test_size=0.1, shuffle=False)\n",
    "    \n",
    "    data_chunks[i] = [X_train, X_test, y_train, y_test]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 298,
   "id": "71d7e950-3312-446c-b945-2b2be2c31c99",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "83"
      ]
     },
     "execution_count": 298,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(data_chunks)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 445,
   "id": "697f1ad0-8f74-4036-bbcb-87483757eb49",
   "metadata": {
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0 0:00:34.882691\n",
      "1 0:00:38.803073\n",
      "2 0:00:39.450899\n",
      "3 0:00:32.226266\n",
      "4 0:00:37.953692\n",
      "5 0:00:34.061250\n",
      "6 0:00:35.191735\n",
      "7 0:00:33.103298\n",
      "8 0:00:38.515957\n",
      "9 0:00:29.165351\n",
      "10 0:00:37.618404\n",
      "11 0:00:33.283527\n",
      "12 0:00:32.248827\n",
      "13 0:00:28.423358\n",
      "14 0:00:30.701219\n",
      "15 0:00:36.321846\n",
      "16 0:00:32.357664\n",
      "17 0:00:43.636373\n",
      "18 0:00:42.098289\n",
      "19 0:00:34.361146\n",
      "20 0:00:29.263793\n",
      "21 0:00:38.014431\n",
      "22 0:00:39.193923\n",
      "23 0:00:33.672841\n",
      "24 0:00:33.079565\n",
      "25 0:00:32.054457\n",
      "26 0:00:36.551696\n",
      "27 0:00:42.581497\n",
      "28 0:00:32.878658\n",
      "29 0:00:30.888633\n",
      "30 0:00:42.599000\n",
      "31 0:00:35.733148\n",
      "32 0:00:33.403454\n",
      "33 0:00:36.339927\n",
      "34 0:00:31.605689\n",
      "35 0:00:27.986178\n",
      "36 0:00:32.113327\n",
      "37 0:00:31.196845\n",
      "38 0:00:27.205261\n",
      "39 0:00:39.608621\n",
      "40 0:00:32.352539\n",
      "41 0:00:38.165709\n",
      "42 0:00:39.408921\n",
      "43 0:00:36.972802\n",
      "44 0:00:39.013656\n",
      "45 0:00:36.829719\n",
      "46 0:00:32.514700\n",
      "47 0:00:40.306100\n",
      "48 0:00:33.911956\n",
      "49 0:00:27.234791\n",
      "50 0:00:35.818384\n",
      "51 0:00:41.222537\n",
      "52 0:00:28.073730\n",
      "53 0:00:37.703285\n",
      "54 0:00:37.342158\n",
      "55 0:00:33.310946\n",
      "56 0:00:38.334503\n",
      "57 0:00:34.051274\n",
      "58 0:00:33.232088\n",
      "59 0:00:32.752133\n",
      "60 0:00:40.871127\n",
      "61 0:00:36.182622\n",
      "62 0:00:36.696790\n",
      "63 0:00:27.591246\n",
      "64 0:00:43.359025\n",
      "65 0:00:34.133266\n",
      "66 0:00:36.356481\n",
      "67 0:00:37.789763\n",
      "68 0:00:32.935878\n",
      "69 0:00:29.618581\n",
      "70 0:00:37.413558\n",
      "71 0:00:28.735129\n",
      "72 0:00:38.142519\n",
      "73 0:00:33.908562\n",
      "74 0:00:27.637586\n",
      "75 0:00:33.108258\n",
      "76 0:00:42.513551\n",
      "77 0:00:33.601048\n",
      "78 0:00:33.902267\n",
      "79 0:00:39.394852\n",
      "80 0:00:27.938742\n",
      "81 0:00:33.763654\n",
      "82 0:00:30.946019\n",
      "CPU times: user 27min 13s, sys: 42.6 s, total: 27min 56s\n",
      "Wall time: 48min 15s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "chunk_results = {}\n",
    "indices = data_chunks.keys()\n",
    "\n",
    "indices = list(indices)\n",
    "\n",
    "mod_list = ['GB', 'RF', 'AB', 'KR', 'R', 'LA', 'EN', 'NW', 'KN']\n",
    "\n",
    "for i in indices:\n",
    "    t = dt.datetime.now()\n",
    "    chunk_results[i] = {}\n",
    "    for ml in mod_list:\n",
    "        chunk_results[i][ml] = model_selection(ml, data_chunks[i])\n",
    "    print(i, dt.datetime.now()-t)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f47908b6-4716-41b5-ace4-e777bf574ea8",
   "metadata": {},
   "outputs": [
    {
     "ename": "",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31mRunning cells with 'FinancialVolatility' requires the ipykernel package.\n",
      "\u001b[1;31mRun the following command to install 'ipykernel' into the Python environment. \n",
      "\u001b[1;31mCommand: 'conda install -n FinancialVolatility ipykernel --update-deps --force-reinstall'"
     ]
    }
   ],
   "source": [
    "# mod_list = ['GB', 'RF', 'AB', 'KR', 'R', 'LA', 'EN', 'NW', 'KN']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8bac0e9b",
   "metadata": {},
   "outputs": [
    {
     "ename": "",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31mRunning cells with 'FinancialVolatility' requires the ipykernel package.\n",
      "\u001b[1;31mRun the following command to install 'ipykernel' into the Python environment. \n",
      "\u001b[1;31mCommand: 'conda install -n FinancialVolatility ipykernel --update-deps --force-reinstall'"
     ]
    }
   ],
   "source": [
    "pip install sklearn==1.0.1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "97e1bb75",
   "metadata": {},
   "outputs": [
    {
     "ename": "",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31mRunning cells with 'FinancialVolatility' requires the ipykernel package.\n",
      "\u001b[1;31mRun the following command to install 'ipykernel' into the Python environment. \n",
      "\u001b[1;31mCommand: 'conda install -n FinancialVolatility ipykernel --update-deps --force-reinstall'"
     ]
    }
   ],
   "source": [
    "conda install -n FinancialVolatility ipykernel --update-deps --force-reinstall"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "3849245d",
   "metadata": {},
   "outputs": [
    {
     "ename": "DistributionNotFound",
     "evalue": "The 'sklearn==1.0.0' distribution was not found and is required by the application",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mDistributionNotFound\u001b[0m                      Traceback (most recent call last)",
      "\u001b[1;32mc:\\Users\\soren\\Documents\\GitHub\\Winter2023\\Nonparametric\\Project\\Forecasting-Realized-Volatility-for-Crude-Oil-Futures-with-High-Frequency-Data\\Final Project (server).ipynb Cell 79\u001b[0m in \u001b[0;36m1\n\u001b[1;32m----> <a href='vscode-notebook-cell:/c%3A/Users/soren/Documents/GitHub/Winter2023/Nonparametric/Project/Forecasting-Realized-Volatility-for-Crude-Oil-Futures-with-High-Frequency-Data/Final%20Project%20%28server%29.ipynb#Y164sZmlsZQ%3D%3D?line=0'>1</a>\u001b[0m pkg_resources\u001b[39m.\u001b[39;49mrequire(\u001b[39m\"\u001b[39;49m\u001b[39msklearn==1.0.0\u001b[39;49m\u001b[39m\"\u001b[39;49m)\n",
      "File \u001b[1;32mc:\\Users\\soren\\AppData\\Local\\R-MINI~1\\envs\\stat-rethink2-pymc\\lib\\site-packages\\pkg_resources\\__init__.py:909\u001b[0m, in \u001b[0;36mWorkingSet.require\u001b[1;34m(self, *requirements)\u001b[0m\n\u001b[0;32m    900\u001b[0m \u001b[39mdef\u001b[39;00m \u001b[39mrequire\u001b[39m(\u001b[39mself\u001b[39m, \u001b[39m*\u001b[39mrequirements):\n\u001b[0;32m    901\u001b[0m     \u001b[39m\"\"\"Ensure that distributions matching `requirements` are activated\u001b[39;00m\n\u001b[0;32m    902\u001b[0m \n\u001b[0;32m    903\u001b[0m \u001b[39m    `requirements` must be a string or a (possibly-nested) sequence\u001b[39;00m\n\u001b[1;32m   (...)\u001b[0m\n\u001b[0;32m    907\u001b[0m \u001b[39m    included, even if they were already activated in this working set.\u001b[39;00m\n\u001b[0;32m    908\u001b[0m \u001b[39m    \"\"\"\u001b[39;00m\n\u001b[1;32m--> 909\u001b[0m     needed \u001b[39m=\u001b[39m \u001b[39mself\u001b[39;49m\u001b[39m.\u001b[39;49mresolve(parse_requirements(requirements))\n\u001b[0;32m    911\u001b[0m     \u001b[39mfor\u001b[39;00m dist \u001b[39min\u001b[39;00m needed:\n\u001b[0;32m    912\u001b[0m         \u001b[39mself\u001b[39m\u001b[39m.\u001b[39madd(dist)\n",
      "File \u001b[1;32mc:\\Users\\soren\\AppData\\Local\\R-MINI~1\\envs\\stat-rethink2-pymc\\lib\\site-packages\\pkg_resources\\__init__.py:795\u001b[0m, in \u001b[0;36mWorkingSet.resolve\u001b[1;34m(self, requirements, env, installer, replace_conflicting, extras)\u001b[0m\n\u001b[0;32m    793\u001b[0m         \u001b[39mif\u001b[39;00m dist \u001b[39mis\u001b[39;00m \u001b[39mNone\u001b[39;00m:\n\u001b[0;32m    794\u001b[0m             requirers \u001b[39m=\u001b[39m required_by\u001b[39m.\u001b[39mget(req, \u001b[39mNone\u001b[39;00m)\n\u001b[1;32m--> 795\u001b[0m             \u001b[39mraise\u001b[39;00m DistributionNotFound(req, requirers)\n\u001b[0;32m    796\u001b[0m     to_activate\u001b[39m.\u001b[39mappend(dist)\n\u001b[0;32m    797\u001b[0m \u001b[39mif\u001b[39;00m dist \u001b[39mnot\u001b[39;00m \u001b[39min\u001b[39;00m req:\n\u001b[0;32m    798\u001b[0m     \u001b[39m# Oops, the \"best\" so far conflicts with a dependency\u001b[39;00m\n",
      "\u001b[1;31mDistributionNotFound\u001b[0m: The 'sklearn==1.0.0' distribution was not found and is required by the application"
     ]
    }
   ],
   "source": [
    "pkg_resources.require(\"sklearn==1.0.0\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "37d7261b-c574-482c-a31d-b4f1821c8d7f",
   "metadata": {},
   "outputs": [
    {
     "ename": "ModuleNotFoundError",
     "evalue": "No module named 'sklearn.neighbors._dist_metrics'",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mModuleNotFoundError\u001b[0m                       Traceback (most recent call last)",
      "\u001b[1;32mc:\\Users\\soren\\Documents\\GitHub\\Winter2023\\Nonparametric\\Project\\Forecasting-Realized-Volatility-for-Crude-Oil-Futures-with-High-Frequency-Data\\Final Project (server).ipynb Cell 79\u001b[0m in \u001b[0;36m7\n\u001b[0;32m      <a href='vscode-notebook-cell:/c%3A/Users/soren/Documents/GitHub/Winter2023/Nonparametric/Project/Forecasting-Realized-Volatility-for-Crude-Oil-Futures-with-High-Frequency-Data/Final%20Project%20%28server%29.ipynb#Y141sZmlsZQ%3D%3D?line=2'>3</a>\u001b[0m \u001b[39m# with open('results.pickle', 'wb') as handle:\u001b[39;00m\n\u001b[0;32m      <a href='vscode-notebook-cell:/c%3A/Users/soren/Documents/GitHub/Winter2023/Nonparametric/Project/Forecasting-Realized-Volatility-for-Crude-Oil-Futures-with-High-Frequency-Data/Final%20Project%20%28server%29.ipynb#Y141sZmlsZQ%3D%3D?line=3'>4</a>\u001b[0m \u001b[39m#     pickle.dump(chunk_results, handle, protocol=pickle.HIGHEST_PROTOCOL)\u001b[39;00m\n\u001b[0;32m      <a href='vscode-notebook-cell:/c%3A/Users/soren/Documents/GitHub/Winter2023/Nonparametric/Project/Forecasting-Realized-Volatility-for-Crude-Oil-Futures-with-High-Frequency-Data/Final%20Project%20%28server%29.ipynb#Y141sZmlsZQ%3D%3D?line=5'>6</a>\u001b[0m \u001b[39mwith\u001b[39;00m \u001b[39mopen\u001b[39m(\u001b[39m'\u001b[39m\u001b[39mresults.pickle\u001b[39m\u001b[39m'\u001b[39m, \u001b[39m'\u001b[39m\u001b[39mrb\u001b[39m\u001b[39m'\u001b[39m) \u001b[39mas\u001b[39;00m handle:\n\u001b[1;32m----> <a href='vscode-notebook-cell:/c%3A/Users/soren/Documents/GitHub/Winter2023/Nonparametric/Project/Forecasting-Realized-Volatility-for-Crude-Oil-Futures-with-High-Frequency-Data/Final%20Project%20%28server%29.ipynb#Y141sZmlsZQ%3D%3D?line=6'>7</a>\u001b[0m     test \u001b[39m=\u001b[39m pickle\u001b[39m.\u001b[39;49mload(handle)\n\u001b[0;32m      <a href='vscode-notebook-cell:/c%3A/Users/soren/Documents/GitHub/Winter2023/Nonparametric/Project/Forecasting-Realized-Volatility-for-Crude-Oil-Futures-with-High-Frequency-Data/Final%20Project%20%28server%29.ipynb#Y141sZmlsZQ%3D%3D?line=8'>9</a>\u001b[0m \u001b[39mprint\u001b[39m(test \u001b[39m==\u001b[39m chunk_results)\n",
      "\u001b[1;31mModuleNotFoundError\u001b[0m: No module named 'sklearn.neighbors._dist_metrics'"
     ]
    }
   ],
   "source": [
    "import pickle\n",
    "\n",
    "# with open('results.pickle', 'wb') as handle:\n",
    "#     pickle.dump(chunk_results, handle, protocol=pickle.HIGHEST_PROTOCOL)\n",
    "\n",
    "with open('results.pickle', 'rb') as handle:\n",
    "    test = pickle.load(handle)\n",
    "\n",
    "print(test == chunk_results)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 446,
   "id": "3260aa7e-f8a5-44d8-87a3-05c828bbfe5b",
   "metadata": {},
   "outputs": [],
   "source": [
    "for label in chunk_results.keys():\n",
    "    chunk_results[label]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 447,
   "id": "d7c8b02f-6367-42e5-b41e-a3cabdb9f768",
   "metadata": {},
   "outputs": [],
   "source": [
    "acc_dict = {}\n",
    "for chunk in chunk_results.keys():\n",
    "    for m in mod_list:\n",
    "        acc_dict[str(chunk)+m] = chunk_results[chunk][m][-1]\n",
    "        acc_dict[str(chunk)+m]['model'] = m"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 474,
   "id": "0d005353-dd46-41e1-8827-66c158b234be",
   "metadata": {},
   "outputs": [],
   "source": [
    "acc_df = pd.DataFrame(acc_dict).transpose()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 477,
   "id": "fe540053-3c9d-4fff-aabc-13fb3cb349fb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>MAE</th>\n",
       "      <th>MAPE</th>\n",
       "      <th>MSE</th>\n",
       "      <th>RMSE</th>\n",
       "      <th>model</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0GB</th>\n",
       "      <td>0.000709911</td>\n",
       "      <td>31.0023</td>\n",
       "      <td>3.2138e-06</td>\n",
       "      <td>0.00179271</td>\n",
       "      <td>Gradient Boosting</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0RF</th>\n",
       "      <td>0.000628076</td>\n",
       "      <td>27.4285</td>\n",
       "      <td>2.89619e-06</td>\n",
       "      <td>0.00170182</td>\n",
       "      <td>RF</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0AB</th>\n",
       "      <td>0.000635671</td>\n",
       "      <td>27.7602</td>\n",
       "      <td>2.83321e-06</td>\n",
       "      <td>0.00168322</td>\n",
       "      <td>AB</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0KR</th>\n",
       "      <td>0.000511138</td>\n",
       "      <td>22.3217</td>\n",
       "      <td>1.92972e-06</td>\n",
       "      <td>0.00138914</td>\n",
       "      <td>KR</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0R</th>\n",
       "      <td>0.000599952</td>\n",
       "      <td>26.2003</td>\n",
       "      <td>2.22889e-06</td>\n",
       "      <td>0.00149295</td>\n",
       "      <td>R</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>82R</th>\n",
       "      <td>8.78483e-05</td>\n",
       "      <td>7.82423</td>\n",
       "      <td>1.26318e-08</td>\n",
       "      <td>0.000112391</td>\n",
       "      <td>R</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>82LA</th>\n",
       "      <td>0.000126671</td>\n",
       "      <td>11.282</td>\n",
       "      <td>2.52705e-08</td>\n",
       "      <td>0.000158967</td>\n",
       "      <td>LA</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>82EN</th>\n",
       "      <td>9.14215e-05</td>\n",
       "      <td>8.14248</td>\n",
       "      <td>1.31909e-08</td>\n",
       "      <td>0.000114852</td>\n",
       "      <td>EN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>82NW</th>\n",
       "      <td>0.000585605</td>\n",
       "      <td>52.157</td>\n",
       "      <td>4.88119e-07</td>\n",
       "      <td>0.000698655</td>\n",
       "      <td>NW</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>82KN</th>\n",
       "      <td>0.000602774</td>\n",
       "      <td>53.6862</td>\n",
       "      <td>5.27404e-07</td>\n",
       "      <td>0.000726226</td>\n",
       "      <td>KN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>747 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "              MAE     MAPE          MSE         RMSE              model\n",
       "0GB   0.000709911  31.0023   3.2138e-06   0.00179271  Gradient Boosting\n",
       "0RF   0.000628076  27.4285  2.89619e-06   0.00170182                 RF\n",
       "0AB   0.000635671  27.7602  2.83321e-06   0.00168322                 AB\n",
       "0KR   0.000511138  22.3217  1.92972e-06   0.00138914                 KR\n",
       "0R    0.000599952  26.2003  2.22889e-06   0.00149295                  R\n",
       "...           ...      ...          ...          ...                ...\n",
       "82R   8.78483e-05  7.82423  1.26318e-08  0.000112391                  R\n",
       "82LA  0.000126671   11.282  2.52705e-08  0.000158967                 LA\n",
       "82EN  9.14215e-05  8.14248  1.31909e-08  0.000114852                 EN\n",
       "82NW  0.000585605   52.157  4.88119e-07  0.000698655                 NW\n",
       "82KN  0.000602774  53.6862  5.27404e-07  0.000726226                 KN\n",
       "\n",
       "[747 rows x 5 columns]"
      ]
     },
     "execution_count": 477,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "acc_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 479,
   "id": "d198c88c-7502-40ba-85f6-3213c78a916d",
   "metadata": {},
   "outputs": [],
   "source": [
    "acc_df.loc[acc_df['model'] == 'GB', 'model'] = 'Gradient Boosting'\n",
    "acc_df.loc[acc_df['model'] == 'AB', 'model'] = 'AdaBoost'\n",
    "acc_df.loc[acc_df['model'] == 'EN', 'model'] = 'Elastic Net'\n",
    "acc_df.loc[acc_df['model'] == 'KN', 'model'] = 'k-Neighbors Regression'\n",
    "acc_df.loc[acc_df['model'] == 'KR', 'model'] = 'Kenel Ridge Regression'\n",
    "acc_df.loc[acc_df['model'] == 'LA', 'model'] = 'Lasso'\n",
    "acc_df.loc[acc_df['model'] == 'NW', 'model'] = 'Nadarya-Watson'\n",
    "acc_df.loc[acc_df['model'] == 'R', 'model'] = 'Ridge'\n",
    "acc_df.loc[acc_df['model'] == 'RF', 'model'] = 'Random Forest'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 484,
   "id": "dfd4d7cb-9290-4aa6-be81-8c641b0aa5d1",
   "metadata": {},
   "outputs": [],
   "source": [
    "acc_df = acc_df.rename(columns={'model':'Model'})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 486,
   "id": "7967c437-9b5a-49b9-907b-d46e10e3001b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Model\n",
       "AdaBoost                  17.912190\n",
       "Elastic Net               15.822203\n",
       "Gradient Boosting         14.627260\n",
       "Kenel Ridge Regression    13.368425\n",
       "Lasso                     17.265219\n",
       "Nadarya-Watson            47.012259\n",
       "Random Forest             14.635831\n",
       "Ridge                     15.543471\n",
       "k-Neighbors Regression    42.037707\n",
       "Name: MAPE, dtype: float64"
      ]
     },
     "execution_count": 486,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "acc_df.reset_index(drop=True).groupby('Model')['MAPE'].apply(lambda x: np.mean(x))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 494,
   "id": "616a4214-ea97-4a67-83cb-3fa064f5f48b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>MAE</th>\n",
       "      <th>MAPE</th>\n",
       "      <th>MSE</th>\n",
       "      <th>RMSE</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Model</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Kenel Ridge Regression</th>\n",
       "      <td>0.000317</td>\n",
       "      <td>13.368425</td>\n",
       "      <td>0.000002</td>\n",
       "      <td>0.000665</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Gradient Boosting</th>\n",
       "      <td>0.000354</td>\n",
       "      <td>14.627260</td>\n",
       "      <td>0.000004</td>\n",
       "      <td>0.000804</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Random Forest</th>\n",
       "      <td>0.000365</td>\n",
       "      <td>14.635831</td>\n",
       "      <td>0.000004</td>\n",
       "      <td>0.000787</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Ridge</th>\n",
       "      <td>0.000367</td>\n",
       "      <td>15.543471</td>\n",
       "      <td>0.000002</td>\n",
       "      <td>0.000709</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Elastic Net</th>\n",
       "      <td>0.000375</td>\n",
       "      <td>15.822203</td>\n",
       "      <td>0.000002</td>\n",
       "      <td>0.000721</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Lasso</th>\n",
       "      <td>0.000396</td>\n",
       "      <td>17.265219</td>\n",
       "      <td>0.000002</td>\n",
       "      <td>0.000747</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AdaBoost</th>\n",
       "      <td>0.000455</td>\n",
       "      <td>17.912190</td>\n",
       "      <td>0.000004</td>\n",
       "      <td>0.000865</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>k-Neighbors Regression</th>\n",
       "      <td>0.000905</td>\n",
       "      <td>42.037707</td>\n",
       "      <td>0.000005</td>\n",
       "      <td>0.001403</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Nadarya-Watson</th>\n",
       "      <td>0.001017</td>\n",
       "      <td>47.012259</td>\n",
       "      <td>0.000005</td>\n",
       "      <td>0.001468</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                             MAE       MAPE       MSE      RMSE\n",
       "Model                                                          \n",
       "Kenel Ridge Regression  0.000317  13.368425  0.000002  0.000665\n",
       "Gradient Boosting       0.000354  14.627260  0.000004  0.000804\n",
       "Random Forest           0.000365  14.635831  0.000004  0.000787\n",
       "Ridge                   0.000367  15.543471  0.000002  0.000709\n",
       "Elastic Net             0.000375  15.822203  0.000002  0.000721\n",
       "Lasso                   0.000396  17.265219  0.000002  0.000747\n",
       "AdaBoost                0.000455  17.912190  0.000004  0.000865\n",
       "k-Neighbors Regression  0.000905  42.037707  0.000005  0.001403\n",
       "Nadarya-Watson          0.001017  47.012259  0.000005  0.001468"
      ]
     },
     "execution_count": 494,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "acc_df.reset_index(drop=True).groupby('Model')[['MAE', 'MAPE', 'MSE', 'RMSE']].apply(lambda x: np.mean(x)).sort_values('MAPE').round(6)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 495,
   "id": "78dd14ff-4a18-43e7-a8f5-e5d0cf084e6b",
   "metadata": {},
   "outputs": [],
   "source": [
    "acc_df.reset_index(drop=True).groupby('Model')[['MAE', 'MAPE', 'MSE', 'RMSE']].apply(lambda x: np.mean(x)).sort_values('MAPE').round(6).to_csv('acc.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 453,
   "id": "504dc844-bd06-4b7d-a74a-44a9f301cd9c",
   "metadata": {},
   "outputs": [],
   "source": [
    "def tree_feat_importance(m, df):\n",
    "    return pd.DataFrame({'cols':df.columns, 'feat_imp':m.feature_importances_}\n",
    "                       ).sort_values('feat_imp', ascending=False)\n",
    "def ensemble_tree_feat_importance(m, df):\n",
    "    feature_importances = np.mean([\n",
    "    tree.feature_importances_ for tree in m.estimators_\n",
    "                                    ], axis=0)\n",
    "    return pd.DataFrame({'cols':df.columns, 'feat_imp':feature_importances}\n",
    "                       ).sort_values('feat_imp', ascending=False)\n",
    "def plot_fi(fi, ax): return ax.barh(fi['cols'], fi['feat_imp'])\n",
    "def xgb_feat_importance(m, df):\n",
    "    feature_importances = m.feature_importances_\n",
    "    return pd.DataFrame({'cols':df.columns, 'feat_imp':feature_importances}\n",
    "                       ).sort_values('feat_imp', ascending=False)\n",
    "\n",
    "def many_ensemble_fi(models, df):\n",
    "    feature_imps = []\n",
    "    # add tree feature importances for all models\n",
    "    for m in models:\n",
    "        feature_imps.append(np.mean([tree.feature_importances_ for tree in m.estimators_], axis=0))\n",
    "    # return np.mean(feature_imps, axis=0)\n",
    "    return pd.DataFrame({'cols':df.columns, 'feat_imp':np.mean(feature_imps, axis=0)}\n",
    "                       ).sort_values('feat_imp', ascending=False)\n",
    "def many_gb_fi(models, df):\n",
    "    '''For GradientBoosting objects'''\n",
    "    feature_imps = []\n",
    "    # add tree feature importances for all models\n",
    "    for m in models:\n",
    "        feature_imps.append(m.feature_importances_)\n",
    "    # return np.mean(feature_imps, axis=0)\n",
    "    return pd.DataFrame({'cols':df.columns, 'feat_imp':np.mean(feature_imps, axis=0)}\n",
    "                       ).sort_values('feat_imp', ascending=False)\n",
    "\n",
    "def many_linear_fi(models, df):\n",
    "    '''For Linear objects'''\n",
    "    feature_imps = []\n",
    "    # add tree feature importances for all models\n",
    "    for m in models:\n",
    "        feature_imps.append(m.coef_)\n",
    "    # return np.mean(feature_imps, axis=0)\n",
    "    return pd.DataFrame({'cols':df.columns, 'feat_imp':np.mean(feature_imps, axis=0)}\n",
    "                       ).sort_values('feat_imp', ascending=False)\n",
    "def many_kr_fi(models, df):\n",
    "    '''For KernelRidge objects'''\n",
    "    feature_imps = []\n",
    "    # add tree feature importances for all models\n",
    "    for m in models:\n",
    "        feature_imps.append(m.coef_)\n",
    "    # return np.mean(feature_imps, axis=0)\n",
    "    return pd.DataFrame({'cols':df.columns, 'feat_imp':np.mean(feature_imps, axis=0)}\n",
    "                       ).sort_values('feat_imp', ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 454,
   "id": "13595ad4-61e3-49f0-b808-c98f9e30a59a",
   "metadata": {
    "tags": []
   },
   "outputs": [],
   "source": [
    "fi_dict = {}\n",
    "fi_dict['RF'] = []\n",
    "fi_dict['GB'] = []\n",
    "fi_dict['AB'] = []\n",
    "fi_dict['EN'] = []\n",
    "fi_dict['KN'] = []\n",
    "fi_dict['KR'] = []\n",
    "fi_dict['LA'] = []\n",
    "fi_dict['NW'] = []\n",
    "fi_dict['R'] = []\n",
    "\n",
    "for m in mod_list:\n",
    "    for chunk in chunk_results.items():\n",
    "        fi_dict[m].append(chunk[1][m][0])\n",
    "        \n",
    "    #ensemble\n",
    "    # if (m == 'RF'):\n",
    "    #     for chunk in chunk_results.items():\n",
    "    #         fi_dict[m].append(chunk[1][m][0])\n",
    "    # #boosting\n",
    "    # elif (m == 'GB') or (m == 'AB'):\n",
    "    #     for chunk in chunk_results.items():\n",
    "    #         fi_dict[m].append(chunk[1][m][0])\n",
    "    # #linear\n",
    "    # elif (m == 'EN'):\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 455,
   "id": "7b19b8d3-98d4-46b1-97ee-c7fb24ec5f5c",
   "metadata": {},
   "outputs": [],
   "source": [
    "fi_mods = list(set(mod_list) - set(['NW', 'KN', 'KR']))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 501,
   "id": "0ba823dc-b0e2-4ad2-8b7a-efc87c62fd20",
   "metadata": {},
   "outputs": [],
   "source": [
    "name_map = {'GB':'Gradient Boosting', 'AB':'AdaBoost', 'EN':'Elastic Net', 'KN':'k-Neighbors Regression', 'KR':'Kernel Ridge Regression', \n",
    "            'LA':'Lasso', 'NW':'Nadarya-Watson', 'R':'Ridge', 'RF':'Random Forest'}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 507,
   "id": "6c7337ec-9bc8-422c-aa14-d916b2b23f30",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2160x1800 with 6 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig,ax = plt.subplots(3,2, figsize=(30,25))\n",
    "\n",
    "k=0\n",
    "for i in range(3):\n",
    "    for j in range(2):\n",
    "        m = fi_mods[k]\n",
    "        # plt.axes(ax[i][j])\n",
    "        ax[i][j].set_title(name_map[m], fontsize=20)\n",
    "        ax[i][j].set_ylabel('Feature Name', fontsize=20)\n",
    "        ax[i][j].set_xlabel('Feature Importance', fontsize=20)\n",
    "        if (m=='RF'):\n",
    "            plot_fi(many_ensemble_fi(fi_dict['RF'],X_train), ax[i][j])\n",
    "        elif (m == 'GB') or (m == 'AB'):\n",
    "            plot_fi(many_gb_fi(fi_dict[m],X_train), ax[i][j])\n",
    "        elif (m == 'R') or (m == 'LA') or (m == 'EN'):\n",
    "            plot_fi(many_linear_fi(fi_dict[m],X_train), ax[i][j])\n",
    "            ax[i][j].set_xlabel('Mean Coefficient', fontsize=20)\n",
    "        \n",
    "        k+=1\n",
    "plt.savefig('feat_imp.png')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d246dac0-3104-4ba3-9030-75062cb0668b",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 534,
   "id": "65509d8f-441a-428f-9d40-fb6354c4b8ad",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Close_fut</th>\n",
       "      <th>Close_gas</th>\n",
       "      <th>Close_btc</th>\n",
       "      <th>Close_bond</th>\n",
       "      <th>12_oil_std</th>\n",
       "      <th>Fut_close_rv</th>\n",
       "      <th>Fut_close_rav</th>\n",
       "      <th>Fut_close_bv</th>\n",
       "      <th>Fut_close_tq</th>\n",
       "      <th>Fut_close_jump_test</th>\n",
       "      <th>Fut_close_rv_m</th>\n",
       "      <th>Fut_close_rv_p</th>\n",
       "      <th>Fut_close_sjv_p</th>\n",
       "      <th>Fut_close_sjv_m</th>\n",
       "      <th>Fut_close_rs</th>\n",
       "      <th>Fut_close_rk</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Close_fut</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.106296</td>\n",
       "      <td>0.123685</td>\n",
       "      <td>-0.143304</td>\n",
       "      <td>-0.023724</td>\n",
       "      <td>0.021084</td>\n",
       "      <td>0.014643</td>\n",
       "      <td>0.111210</td>\n",
       "      <td>0.111583</td>\n",
       "      <td>-0.028135</td>\n",
       "      <td>-0.057840</td>\n",
       "      <td>0.153223</td>\n",
       "      <td>0.099861</td>\n",
       "      <td>0.183715</td>\n",
       "      <td>0.161040</td>\n",
       "      <td>-0.005454</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Close_gas</th>\n",
       "      <td>0.106296</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.059274</td>\n",
       "      <td>-0.035915</td>\n",
       "      <td>0.000776</td>\n",
       "      <td>-0.002852</td>\n",
       "      <td>-0.000330</td>\n",
       "      <td>-0.002882</td>\n",
       "      <td>-0.001872</td>\n",
       "      <td>-0.026197</td>\n",
       "      <td>-0.004707</td>\n",
       "      <td>0.000865</td>\n",
       "      <td>0.009143</td>\n",
       "      <td>0.004213</td>\n",
       "      <td>0.029669</td>\n",
       "      <td>0.015411</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Close_btc</th>\n",
       "      <td>0.123685</td>\n",
       "      <td>0.059274</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.071214</td>\n",
       "      <td>-0.010462</td>\n",
       "      <td>-0.005936</td>\n",
       "      <td>-0.004143</td>\n",
       "      <td>0.002291</td>\n",
       "      <td>0.001701</td>\n",
       "      <td>-0.021023</td>\n",
       "      <td>-0.011055</td>\n",
       "      <td>0.003967</td>\n",
       "      <td>0.004120</td>\n",
       "      <td>0.018767</td>\n",
       "      <td>0.042610</td>\n",
       "      <td>0.001366</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Close_bond</th>\n",
       "      <td>-0.143304</td>\n",
       "      <td>-0.035915</td>\n",
       "      <td>-0.071214</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.021569</td>\n",
       "      <td>0.015148</td>\n",
       "      <td>0.009424</td>\n",
       "      <td>-0.001608</td>\n",
       "      <td>-0.001185</td>\n",
       "      <td>0.013096</td>\n",
       "      <td>0.024794</td>\n",
       "      <td>-0.004233</td>\n",
       "      <td>-0.006403</td>\n",
       "      <td>-0.039113</td>\n",
       "      <td>-0.043156</td>\n",
       "      <td>0.003737</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12_oil_std</th>\n",
       "      <td>-0.023724</td>\n",
       "      <td>0.000776</td>\n",
       "      <td>-0.010462</td>\n",
       "      <td>0.021569</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.748941</td>\n",
       "      <td>0.928185</td>\n",
       "      <td>0.596931</td>\n",
       "      <td>0.510300</td>\n",
       "      <td>0.013988</td>\n",
       "      <td>0.724373</td>\n",
       "      <td>0.654717</td>\n",
       "      <td>0.367344</td>\n",
       "      <td>-0.679001</td>\n",
       "      <td>-0.022689</td>\n",
       "      <td>0.279652</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Fut_close_rv</th>\n",
       "      <td>0.021084</td>\n",
       "      <td>-0.002852</td>\n",
       "      <td>-0.005936</td>\n",
       "      <td>0.015148</td>\n",
       "      <td>0.748941</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.711486</td>\n",
       "      <td>0.927591</td>\n",
       "      <td>0.874798</td>\n",
       "      <td>0.001407</td>\n",
       "      <td>0.962145</td>\n",
       "      <td>0.882894</td>\n",
       "      <td>0.205099</td>\n",
       "      <td>-0.773536</td>\n",
       "      <td>-0.020548</td>\n",
       "      <td>0.075247</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Fut_close_rav</th>\n",
       "      <td>0.014643</td>\n",
       "      <td>-0.000330</td>\n",
       "      <td>-0.004143</td>\n",
       "      <td>0.009424</td>\n",
       "      <td>0.928185</td>\n",
       "      <td>0.711486</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.626256</td>\n",
       "      <td>0.522963</td>\n",
       "      <td>0.003772</td>\n",
       "      <td>0.659051</td>\n",
       "      <td>0.672105</td>\n",
       "      <td>0.347983</td>\n",
       "      <td>-0.555482</td>\n",
       "      <td>-0.019540</td>\n",
       "      <td>0.136189</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Fut_close_bv</th>\n",
       "      <td>0.111210</td>\n",
       "      <td>-0.002882</td>\n",
       "      <td>0.002291</td>\n",
       "      <td>-0.001608</td>\n",
       "      <td>0.596931</td>\n",
       "      <td>0.927591</td>\n",
       "      <td>0.626256</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.965960</td>\n",
       "      <td>-0.000394</td>\n",
       "      <td>0.842626</td>\n",
       "      <td>0.904858</td>\n",
       "      <td>0.088625</td>\n",
       "      <td>-0.522892</td>\n",
       "      <td>-0.008010</td>\n",
       "      <td>0.016180</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Fut_close_tq</th>\n",
       "      <td>0.111583</td>\n",
       "      <td>-0.001872</td>\n",
       "      <td>0.001701</td>\n",
       "      <td>-0.001185</td>\n",
       "      <td>0.510300</td>\n",
       "      <td>0.874798</td>\n",
       "      <td>0.522963</td>\n",
       "      <td>0.965960</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.000134</td>\n",
       "      <td>0.797047</td>\n",
       "      <td>0.849262</td>\n",
       "      <td>0.031411</td>\n",
       "      <td>-0.478920</td>\n",
       "      <td>-0.007132</td>\n",
       "      <td>0.010209</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Fut_close_jump_test</th>\n",
       "      <td>-0.028135</td>\n",
       "      <td>-0.026197</td>\n",
       "      <td>-0.021023</td>\n",
       "      <td>0.013096</td>\n",
       "      <td>0.013988</td>\n",
       "      <td>0.001407</td>\n",
       "      <td>0.003772</td>\n",
       "      <td>-0.000394</td>\n",
       "      <td>-0.000134</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.001988</td>\n",
       "      <td>0.000151</td>\n",
       "      <td>0.000906</td>\n",
       "      <td>-0.003285</td>\n",
       "      <td>-0.010638</td>\n",
       "      <td>0.038370</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Fut_close_rv_m</th>\n",
       "      <td>-0.057840</td>\n",
       "      <td>-0.004707</td>\n",
       "      <td>-0.011055</td>\n",
       "      <td>0.024794</td>\n",
       "      <td>0.724373</td>\n",
       "      <td>0.962145</td>\n",
       "      <td>0.659051</td>\n",
       "      <td>0.842626</td>\n",
       "      <td>0.797047</td>\n",
       "      <td>0.001988</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.721496</td>\n",
       "      <td>0.015807</td>\n",
       "      <td>-0.897047</td>\n",
       "      <td>-0.058598</td>\n",
       "      <td>0.071277</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Fut_close_rv_p</th>\n",
       "      <td>0.153223</td>\n",
       "      <td>0.000865</td>\n",
       "      <td>0.003967</td>\n",
       "      <td>-0.004233</td>\n",
       "      <td>0.654717</td>\n",
       "      <td>0.882894</td>\n",
       "      <td>0.672105</td>\n",
       "      <td>0.904858</td>\n",
       "      <td>0.849262</td>\n",
       "      <td>0.000151</td>\n",
       "      <td>0.721496</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.493848</td>\n",
       "      <td>-0.419691</td>\n",
       "      <td>0.048757</td>\n",
       "      <td>0.068366</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Fut_close_sjv_p</th>\n",
       "      <td>0.099861</td>\n",
       "      <td>0.009143</td>\n",
       "      <td>0.004120</td>\n",
       "      <td>-0.006403</td>\n",
       "      <td>0.367344</td>\n",
       "      <td>0.205099</td>\n",
       "      <td>0.347983</td>\n",
       "      <td>0.088625</td>\n",
       "      <td>0.031411</td>\n",
       "      <td>0.000906</td>\n",
       "      <td>0.015807</td>\n",
       "      <td>0.493848</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.002840</td>\n",
       "      <td>0.119847</td>\n",
       "      <td>0.130809</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Fut_close_sjv_m</th>\n",
       "      <td>0.183715</td>\n",
       "      <td>0.004213</td>\n",
       "      <td>0.018767</td>\n",
       "      <td>-0.039113</td>\n",
       "      <td>-0.679001</td>\n",
       "      <td>-0.773536</td>\n",
       "      <td>-0.555482</td>\n",
       "      <td>-0.522892</td>\n",
       "      <td>-0.478920</td>\n",
       "      <td>-0.003285</td>\n",
       "      <td>-0.897047</td>\n",
       "      <td>-0.419691</td>\n",
       "      <td>0.002840</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.083747</td>\n",
       "      <td>-0.102315</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Fut_close_rs</th>\n",
       "      <td>0.161040</td>\n",
       "      <td>0.029669</td>\n",
       "      <td>0.042610</td>\n",
       "      <td>-0.043156</td>\n",
       "      <td>-0.022689</td>\n",
       "      <td>-0.020548</td>\n",
       "      <td>-0.019540</td>\n",
       "      <td>-0.008010</td>\n",
       "      <td>-0.007132</td>\n",
       "      <td>-0.010638</td>\n",
       "      <td>-0.058598</td>\n",
       "      <td>0.048757</td>\n",
       "      <td>0.119847</td>\n",
       "      <td>0.083747</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.036563</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Fut_close_rk</th>\n",
       "      <td>-0.005454</td>\n",
       "      <td>0.015411</td>\n",
       "      <td>0.001366</td>\n",
       "      <td>0.003737</td>\n",
       "      <td>0.279652</td>\n",
       "      <td>0.075247</td>\n",
       "      <td>0.136189</td>\n",
       "      <td>0.016180</td>\n",
       "      <td>0.010209</td>\n",
       "      <td>0.038370</td>\n",
       "      <td>0.071277</td>\n",
       "      <td>0.068366</td>\n",
       "      <td>0.130809</td>\n",
       "      <td>-0.102315</td>\n",
       "      <td>0.036563</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                     Close_fut  Close_gas  Close_btc  Close_bond  12_oil_std  \\\n",
       "Close_fut             1.000000   0.106296   0.123685   -0.143304   -0.023724   \n",
       "Close_gas             0.106296   1.000000   0.059274   -0.035915    0.000776   \n",
       "Close_btc             0.123685   0.059274   1.000000   -0.071214   -0.010462   \n",
       "Close_bond           -0.143304  -0.035915  -0.071214    1.000000    0.021569   \n",
       "12_oil_std           -0.023724   0.000776  -0.010462    0.021569    1.000000   \n",
       "Fut_close_rv          0.021084  -0.002852  -0.005936    0.015148    0.748941   \n",
       "Fut_close_rav         0.014643  -0.000330  -0.004143    0.009424    0.928185   \n",
       "Fut_close_bv          0.111210  -0.002882   0.002291   -0.001608    0.596931   \n",
       "Fut_close_tq          0.111583  -0.001872   0.001701   -0.001185    0.510300   \n",
       "Fut_close_jump_test  -0.028135  -0.026197  -0.021023    0.013096    0.013988   \n",
       "Fut_close_rv_m       -0.057840  -0.004707  -0.011055    0.024794    0.724373   \n",
       "Fut_close_rv_p        0.153223   0.000865   0.003967   -0.004233    0.654717   \n",
       "Fut_close_sjv_p       0.099861   0.009143   0.004120   -0.006403    0.367344   \n",
       "Fut_close_sjv_m       0.183715   0.004213   0.018767   -0.039113   -0.679001   \n",
       "Fut_close_rs          0.161040   0.029669   0.042610   -0.043156   -0.022689   \n",
       "Fut_close_rk         -0.005454   0.015411   0.001366    0.003737    0.279652   \n",
       "\n",
       "                     Fut_close_rv  Fut_close_rav  Fut_close_bv  Fut_close_tq  \\\n",
       "Close_fut                0.021084       0.014643      0.111210      0.111583   \n",
       "Close_gas               -0.002852      -0.000330     -0.002882     -0.001872   \n",
       "Close_btc               -0.005936      -0.004143      0.002291      0.001701   \n",
       "Close_bond               0.015148       0.009424     -0.001608     -0.001185   \n",
       "12_oil_std               0.748941       0.928185      0.596931      0.510300   \n",
       "Fut_close_rv             1.000000       0.711486      0.927591      0.874798   \n",
       "Fut_close_rav            0.711486       1.000000      0.626256      0.522963   \n",
       "Fut_close_bv             0.927591       0.626256      1.000000      0.965960   \n",
       "Fut_close_tq             0.874798       0.522963      0.965960      1.000000   \n",
       "Fut_close_jump_test      0.001407       0.003772     -0.000394     -0.000134   \n",
       "Fut_close_rv_m           0.962145       0.659051      0.842626      0.797047   \n",
       "Fut_close_rv_p           0.882894       0.672105      0.904858      0.849262   \n",
       "Fut_close_sjv_p          0.205099       0.347983      0.088625      0.031411   \n",
       "Fut_close_sjv_m         -0.773536      -0.555482     -0.522892     -0.478920   \n",
       "Fut_close_rs            -0.020548      -0.019540     -0.008010     -0.007132   \n",
       "Fut_close_rk             0.075247       0.136189      0.016180      0.010209   \n",
       "\n",
       "                     Fut_close_jump_test  Fut_close_rv_m  Fut_close_rv_p  \\\n",
       "Close_fut                      -0.028135       -0.057840        0.153223   \n",
       "Close_gas                      -0.026197       -0.004707        0.000865   \n",
       "Close_btc                      -0.021023       -0.011055        0.003967   \n",
       "Close_bond                      0.013096        0.024794       -0.004233   \n",
       "12_oil_std                      0.013988        0.724373        0.654717   \n",
       "Fut_close_rv                    0.001407        0.962145        0.882894   \n",
       "Fut_close_rav                   0.003772        0.659051        0.672105   \n",
       "Fut_close_bv                   -0.000394        0.842626        0.904858   \n",
       "Fut_close_tq                   -0.000134        0.797047        0.849262   \n",
       "Fut_close_jump_test             1.000000        0.001988        0.000151   \n",
       "Fut_close_rv_m                  0.001988        1.000000        0.721496   \n",
       "Fut_close_rv_p                  0.000151        0.721496        1.000000   \n",
       "Fut_close_sjv_p                 0.000906        0.015807        0.493848   \n",
       "Fut_close_sjv_m                -0.003285       -0.897047       -0.419691   \n",
       "Fut_close_rs                   -0.010638       -0.058598        0.048757   \n",
       "Fut_close_rk                    0.038370        0.071277        0.068366   \n",
       "\n",
       "                     Fut_close_sjv_p  Fut_close_sjv_m  Fut_close_rs  \\\n",
       "Close_fut                   0.099861         0.183715      0.161040   \n",
       "Close_gas                   0.009143         0.004213      0.029669   \n",
       "Close_btc                   0.004120         0.018767      0.042610   \n",
       "Close_bond                 -0.006403        -0.039113     -0.043156   \n",
       "12_oil_std                  0.367344        -0.679001     -0.022689   \n",
       "Fut_close_rv                0.205099        -0.773536     -0.020548   \n",
       "Fut_close_rav               0.347983        -0.555482     -0.019540   \n",
       "Fut_close_bv                0.088625        -0.522892     -0.008010   \n",
       "Fut_close_tq                0.031411        -0.478920     -0.007132   \n",
       "Fut_close_jump_test         0.000906        -0.003285     -0.010638   \n",
       "Fut_close_rv_m              0.015807        -0.897047     -0.058598   \n",
       "Fut_close_rv_p              0.493848        -0.419691      0.048757   \n",
       "Fut_close_sjv_p             1.000000         0.002840      0.119847   \n",
       "Fut_close_sjv_m             0.002840         1.000000      0.083747   \n",
       "Fut_close_rs                0.119847         0.083747      1.000000   \n",
       "Fut_close_rk                0.130809        -0.102315      0.036563   \n",
       "\n",
       "                     Fut_close_rk  \n",
       "Close_fut               -0.005454  \n",
       "Close_gas                0.015411  \n",
       "Close_btc                0.001366  \n",
       "Close_bond               0.003737  \n",
       "12_oil_std               0.279652  \n",
       "Fut_close_rv             0.075247  \n",
       "Fut_close_rav            0.136189  \n",
       "Fut_close_bv             0.016180  \n",
       "Fut_close_tq             0.010209  \n",
       "Fut_close_jump_test      0.038370  \n",
       "Fut_close_rv_m           0.071277  \n",
       "Fut_close_rv_p           0.068366  \n",
       "Fut_close_sjv_p          0.130809  \n",
       "Fut_close_sjv_m         -0.102315  \n",
       "Fut_close_rs             0.036563  \n",
       "Fut_close_rk             1.000000  "
      ]
     },
     "execution_count": 534,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.corr()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "FinancialVolatility",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.16"
  },
  "toc-autonumbering": true,
  "toc-showcode": false,
  "toc-showmarkdowntxt": false,
  "toc-showtags": false
 },
 "nbformat": 4,
 "nbformat_minor": 5
}