finmlkit.feature.core.ma module

finmlkit.feature.core.ma.ewma(y: ndarray[tuple[int, ...], dtype[_ScalarType_co]], span: int) → ndarray[tuple[int, ...], dtype[float64]][source]

Exponentially weighted moving average (EWMA) of a one-dimensional numpy array. Calculates the equivalent of pandas.DataFrame.ewm(…).mean() with adjust=True.

By using this weighting scheme, the function provides a more accurate and unbiased estimate of the EWMA, especially in the early stages of the data series.

Parameters:

  • y – A one-dimensional numpy array of floats.

  • span – The decay window, or ‘span’. Determines how many past points meaningfully impact the EWMA value.

Returns:

The EWMA vector, same length and shape as y.

Raises:

ValueError – If span is less than 1.

Note

This function adjusts for small sample sizes by dividing by the cumulative weight. For more information, see: https://terbe.dev/blog/posts/exponentially-weighted-moving-average

finmlkit.feature.core.ma.sma(array: ndarray[tuple[int, ...], dtype[float64]], window: int) → ndarray[tuple[int, ...], dtype[float64]][source]

Calculate Simple Moving Average (SMA) with Numba for better performance.

Parameters:

  • array – np.array, input array to calculate SMA

  • window – int, window for the rolling average

Returns:

np.array, SMA values