Class SeriesWeightedMovingFeature

• Jump right in for a hands-on

Import

from NitroFE import SeriesWeightedMovingFeature

Series Weighted Moving Feature is a moving window variation and is calculated as

\[ Series\_Weighted\_Moving\_Feature[t] = \frac{\operatorname{operation} (dataframe[t-'lookback\_period' \to t] * dataframe\_for\_weight[t-'lookback\_period' \to t])}{\operatorname{operation} (dataframe\_for\_weight[t-'lookback\_period' \to t])} \]

Methods

__init__(self, lookback_period=4, min_periods=None, operation=None, operation_args=()) special

Parameters:

Name	Type	Description	Default
lookback_period	int	Size of the rolling window for lookback, by default 4	4
min_periods	int	Minimum number of observations in window required to have a value, by default None	None
operation	Callable	operation to perform over values. If None, numpy mean is used, by default None	None
operation_args	tuple	additional agrument values to be sent for operation function	()

Source code in nitrofe\time_based_features\indicator_features\_seriesweightedmovingfeature.py

def __init__(
    self,
    lookback_period: int = 4,
    min_periods: int = None,
    operation: Callable = None,
    operation_args: tuple = (),
):
    """
    Parameters
    ----------
    lookback_period : int, optional
        Size of the rolling window for lookback, by default 4
    min_periods : int, optional
        Minimum number of observations in window required to have a value, by default None
    operation : Callable, optional
        operation to perform over values. If None, numpy mean is used, by default None
    operation_args : tuple, optional
        additional agrument values to be sent for operation function
    """
    self.lookback_period = lookback_period
    self.min_periods = min_periods
    self.operation = np.mean if operation == None else operation
    self.operation_args = operation_args

fit(self, dataframe, dataframe_for_weight, first_fit=True)

For your training/initial fit phase (very first fit) use fit_first=True, and for any production/test implementation pass fit_first=False

Parameters:

Name	Type	Description	Default
dataframe	Union[pandas.core.frame.DataFrame, pandas.core.series.Series]	dataframe containing column values	required
dataframe_for_weight	Union[pandas.core.frame.DataFrame, pandas.core.series.Series]	dataframe containing column values	required
first_fit	bool	Indicator features require past values for calculation. Use True, when calculating for training data (very first fit) Use False, when calculating for subsequent testing/production data { in which case the values, which were saved during the last phase, will be utilized for calculation }, by default True	True

Source code in nitrofe\time_based_features\indicator_features\_seriesweightedmovingfeature.py

def fit(
    self,
    dataframe: Union[pd.DataFrame, pd.Series],
    dataframe_for_weight: Union[pd.DataFrame, pd.Series],
    first_fit: bool = True,
):
    """
    For your training/initial fit phase (very first fit) use fit_first=True, and for any production/test implementation pass fit_first=False

    Parameters
    ----------
    dataframe : Union[pd.DataFrame, pd.Series]
        dataframe containing column values
    dataframe_for_weight : Union[pd.DataFrame, pd.Series]
        dataframe containing column values
    first_fit : bool, optional
        Indicator features require past values for calculation.
        Use True, when calculating for training data (very first fit)
        Use False, when calculating for subsequent testing/production data { in which case the values, which
        were saved during the last phase, will be utilized for calculation }, by default True

    """
    if first_fit:
        self._multiplication_object = weighted_window_features()
        self._weight_object = weighted_window_features()

    if isinstance(dataframe, pd.Series):
        dataframe = dataframe.to_frame()
    if isinstance(dataframe_for_weight, pd.Series):
        dataframe_for_weight = dataframe_for_weight.to_frame()

    multiplication_res = pd.DataFrame(
        np.multiply(dataframe.values, dataframe_for_weight.values),
        columns=dataframe.columns,
    )

    _multiplication_value = (
        self._multiplication_object._template_feature_calculation(
            function_name="_multiplication_object",
            win_function=_identity_window,
            first_fit=first_fit,
            dataframe=multiplication_res,
            window=self.lookback_period,
            min_periods=self.min_periods,
            symmetric=None,
            operation=self.operation,
            operation_args=self.operation_args,
        )
    )

    _weight_value = self._multiplication_object._template_feature_calculation(
        function_name="_weight_value",
        win_function=_identity_window,
        first_fit=first_fit,
        dataframe=dataframe_for_weight,
        window=self.lookback_period,
        min_periods=self.min_periods,
        symmetric=None,
        operation=self.operation,
        operation_args=self.operation_args,
    )

    res = pd.DataFrame(
        (_multiplication_value.values / _weight_value.values),
        columns=dataframe.columns,
    )

    return res