roll_standardize¶

Defined in fynance.features.scale

roll_standardize(X, w=None, a=0, b=1, axis=0, kind_moment='s')[source]

Substitutes the rolling mean and divid by the rolling standard dev.

\[RollStandardize(X)^w_t = b \times \frac{X_t - RollMean(X)^w_t} {RollStd(X)^w_t} + a\]

Parameters:

Xnp.ndarray[dtype, ndim=1 or 2]

Data to scale.

wint, optional

Size of the lagged window of the moving average/standard deviation, must be positive. If w is None or w=0, then w=X.shape[axis]. Default is None.

a, bfloat or array_like, optional

Respectively an additional and multiply factor.

axisint, optional

Axis along which to scale the data.

kind_momentstr {“s”, “w”, “e”}, optional

If “s” (default) then compute basic moving averages and standard deviations, see sma and smstd.
If “w” then compute the weighted moving averages and standard deviations, see wma and wmstd.
If “e” then compute the exponential moving averages and standard deviations, see ema and emstd.

Returns:

np.ndarray[dtype, ndim=1 or 2]: The scaled data.