Replace histogram feature implementation to histogram mode.

fraunhoferportugal · Sep 11, 2024 · 69be8f0 · 69be8f0
1 parent 1d25d5e
commit 69be8f0
Show file tree

Hide file tree

Showing 5 changed files with 43 additions and 58 deletions.
diff --git a/tests/test_calc_features.py b/tests/test_calc_features.py
@@ -75,7 +75,7 @@ def test_input_list_window_multi_axis_multi(self):
         )
         np.testing.assert_array_equal(
             features0.shape,
-            (16, 495),
+            (16, 468),
         )
 
     def test_input_array_window_tosplit_axis_single(self):
@@ -92,7 +92,7 @@ def test_input_array_window_tosplit_axis_single(self):
 
         np.testing.assert_array_equal(
             features1.shape,
-            (16, 165),
+            (16, 156),
         )
 
     def test_input_series_window_tosplit_axis_single(self):
@@ -108,7 +108,7 @@ def test_input_series_window_tosplit_axis_single(self):
 
         np.testing.assert_array_equal(
             features2.shape,
-            (16, 165),
+            (16, 156),
         )
 
     def test_input_dataframe_window_single_axis_multi(self):
@@ -152,7 +152,7 @@ def test_input_series_window_single_axis_single(self):
 
         np.testing.assert_array_equal(
             features5.shape,
-            (1, 40),
+            (1, 31),
         )
 
     def test_input_array_window_single_axis_single(self):
@@ -183,14 +183,14 @@ def test_personal_features(self):
 
         np.testing.assert_array_equal(
             features7.shape,
-            (1, 169),
+            (1, 160),
         )
 
     def test_get_number_features(self):
         feature_sets_size = [get_number_features(get_features_by_domain(domain)) for domain in domains]
         np.testing.assert_array_equal(
             feature_sets_size,
-            [40, 14, 124, 6],  # 184 is the total
+            [31, 14, 124, 6],  # 184 is the total
         )
 
     def test_dataset_extractor(self):

diff --git a/tests/test_features.py b/tests/test_features.py
@@ -34,38 +34,42 @@ class TestFeatures(unittest.TestCase):
     # ############################################### STATISTICAL FEATURES ############################################### #
     def test_hist(self):
         np.testing.assert_almost_equal(
-            hist(const0, 10, 5)["values"],
-            (0.0, 0.0, 0.0, 0.0, 0.0, 20.0, 0.0, 0.0, 0.0, 0.0),
+            hist(const0, 10),
+            0.050000000000000044,
+            decimal=5,
         )
         np.testing.assert_almost_equal(
-            hist(const1, 10, 5)["values"],
-            (0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 20.0, 0.0, 0.0, 0.0),
+            hist(const1, 10),
+            1.05,
+            decimal=5,
         )
         np.testing.assert_almost_equal(
-            hist(constNeg, 10, 5)["values"],
-            (0.0, 0.0, 0.0, 0.0, 20.0, 0.0, 0.0, 0.0, 0.0, 0.0),
+            hist(constNeg, 10),
+            -0.95,
+            decimal=5,
         )
         np.testing.assert_almost_equal(
-            hist(constF, 10, 5)["values"],
-            (0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 20.0, 0.0, 0.0),
+            hist(constF, 10),
+            2.55,
+            decimal=5,
         )
         np.testing.assert_almost_equal(
-            hist(lin, 10, 5)["values"],
-            (0.0, 0.0, 0.0, 0.0, 0.0, 1.0, 1.0, 1.0, 1.0, 2),
+            hist(lin, 10),
+            0.95,
         )
         np.testing.assert_almost_equal(
-            hist(wave, 10, 5)["values"],
-            (0.0, 0.0, 0.0, 0.0, 499, 496, 5, 0.0, 0.0, 0.0),
+            hist(wave, 10),
+            -0.9,
             decimal=5,
         )
         np.testing.assert_almost_equal(
-            hist(offsetWave, 10, 5)["values"],
-            (0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 499, 496, 5, 0.0),
+            hist(offsetWave, 10),
+            1.1,
             decimal=5,
         )
         np.testing.assert_almost_equal(
-            hist(noiseWave, 10, 5)["values"],
-            (0.0, 0.0, 0.0, 48, 446, 450, 56, 0.0, 0.0, 0.0),
+            hist(noiseWave, 10),
+            -0.8862517157830269,
             decimal=5,
         )
 

diff --git a/tests/tests_tools/test_features.json b/tests/tests_tools/test_features.json
@@ -430,14 +430,13 @@
             "tag": "eeg",
             "use": "yes"
         },
-        "Histogram": {
+        "Histogram mode": {
             "complexity": "log",
-            "description": "Computes histogram of the signal.",
+            "description": "Computes the mode of the signal's histogram.",
             "function": "tsfel.hist",
-            "n_features": "nbins",
+            "n_features": 1,
             "parameters": {
-                "nbins": 10,
-                "r": 1
+                "nbins": 10
             },
             "use": "yes"
         },
@@ -704,4 +703,4 @@
             "use": "yes"
         }
     }
-}
+}
diff --git a/tsfel/feature_extraction/features.json b/tsfel/feature_extraction/features.json
@@ -355,15 +355,14 @@
       "use": "yes",
       "tag": "eeg"
     },
-    "Histogram": {
+    "Histogram mode": {
       "complexity": "log",
-      "description": "Computes histogram of the signal.",
+      "description": "Computes the mode of the signal's histogram.",
       "function": "tsfel.hist",
       "parameters": {
-        "nbins": 10,
-        "r": 1
+        "nbins": 10
       },
-      "n_features": "nbins",
+      "n_features": 1,
       "use": "yes"
     },
     "Interquartile range": {

diff --git a/tsfel/feature_extraction/features.py b/tsfel/feature_extraction/features.py
@@ -13,9 +13,6 @@
     + " data points."
 )
 
-future_warn_flag = False
-warnings.simplefilter("once", FutureWarning)
-
 # ############################################# TEMPORAL DOMAIN ##################################################### #
 
 
@@ -473,8 +470,8 @@ def entropy(signal, prob="standard"):
 
 
 @set_domain("domain", "statistical")
-def hist(signal, nbins=10, r=1):
-    """Computes histogram of the signal.
+def hist(signal, nbins=10):
+    """Computes the mode of the signal's histogram.
 
     Feature computational cost: 1
 
@@ -484,34 +481,20 @@ def hist(signal, nbins=10, r=1):
         Input from histogram is computed
     nbins : int
         The number of equal-width bins in the given range
-    r : float
-        The lower(-r) and upper(r) range of the bins
 
     Returns
     -------
-    nd-array
-        The values of the histogram
+    float
+        The mode of the histogram
     """
-    global future_warn_flag
 
-    if not future_warn_flag:
-        warnings.warn(
-            (
-                "The histogram feature was deprecated in version 0.1.8 and will be replaced by the mode of histogram in 0.1.9."
-                " From then on, only a single feature value will be returned."
-            ),
-            FutureWarning,
-        )
-        future_warn_flag = True
+    histsig, bin_edges = np.histogram(signal, bins=nbins)
 
-    # TODO: r value must be revised!
-    histsig, bin_edges = np.histogram(signal, bins=nbins, range=[-r, r])
+    idx_max = np.argmax(histsig)
 
-    names = [
-        str(np.around(bin_edges[i], 2)) + ":" + str(np.around(bin_edges[i + 1], 2)) for i in range(len(bin_edges) - 1)
-    ]
+    mode_hist = (bin_edges[idx_max] + bin_edges[idx_max + 1]) * 0.5
 
-    return {"names": names, "values": histsig}
+    return mode_hist
 
 
 @set_domain("domain", "statistical")