config.json

{
    "models": [
        {
            "enabled": true,
            "model_name": "Cases",
            "model": {
                "type": "neural_net",
                "alpha": 1e-6,
                "hidden_layer_sizes": "auto",
                "learning_rate_init": 0.0008,
                "max_iter": 50000,
                "batch_size": 32,
                "tol": 1e-6,
                "n_iter_no_change": 250,
                "verbose": false
            },
            "datagrabber_class": "CasesDataGrabber",
            "grab_data_from_server": true,
            "offline_dataset_date": "2020-09-26",
            "days_to_predict": 10
        },
        {
            "enabled": true,
            "model_name": "Deaths",
            "model": {
                "type": "neural_net",
                "alpha": 1e-6,
                "hidden_layer_sizes": "auto",
                "learning_rate_init": 0.0002,
                "max_iter": 50000,
                "batch_size": 32,
                "tol": 1e-6,
                "n_iter_no_change": 250,
                "verbose": false
            },
            "datagrabber_class": "DeathsDataGrabber",
            "grab_data_from_server": true,
            "offline_dataset_date": "2020-09-26",
            "days_to_predict": 10
        },
        {
            "_comment": "This is an example of how to use a neural network model with custom hidden layer sizes (instead of 'auto').",
            "enabled": false,
            "model_name": "Deaths using Neural Network with custom hidden layer sizes",
            "model": {
                "_comment": "You may use any valid sklearn's MLPRegressor arguments here.",
                "type": "neural_net",
                "alpha": 1e-5,
                "hidden_layer_sizes": [48, 48],
                "learning_rate_init": 0.008,
                "max_iter": 50000,
                "batch_size": 32,
                "verbose": false
            },
            "datagrabber_class": "DeathsDataGrabber",
            "grab_data_from_server": true,
            "offline_dataset_date": "2020-09-25",
            "days_to_predict": 10
        },
        {
            "_comment": "This is an example of how to use a polynomial regression model which can be useful in most cases.",
            "enabled": false,
            "model_name": "Cases using Polynomial Regression",
            "model": {
                "type": "regression",
                "polynomial_degree": 2
            },
            "datagrabber_class": "CasesDataGrabber",
            "grab_data_from_server": true,
            "offline_dataset_date": "2020-05-15",
            "days_to_predict": 10
        }
    ]
}