paddlets.transform.statistical

class StatsTransform(cols: Union[str, List], start: int = 0, end: int = 1, statistics: List = ['median', 'mean', 'max', 'min', 'std'])[source]

Bases: BaseTransform

Statistical features: ‘median’, ‘mean’, ‘max’, ‘min’, ‘std’

Parameters

cols (str|List) – Name of columns to transform.
start (int) – Start coordinates.
end (int) – End coordinates.
statistics (str|List) – Indicators that need to be counted, default=[‘median’, ‘mean’, ‘max’, ‘min’, ‘std’].

Returns

None

Examples

Given X:
    X
    1
    2
    3
    4

statistics = ['mean'], start = 0, end = 2

after transform:
    X X_mean
    1 nan
    2 1.5
    3 2.5
    4 3.5
Remark: since the first element(1) start index has no value, the result of calculating the mean is nan

statistics = ['mean'], start = 1, end = 3

after transform:
    X X_mean
    1 nan
    2 nan
    3 1.5
    4 2.5

fit_one(tsdata: TSDataset)[source]

Fit the StatsTransform to dataset.

Parameters: tsdata (TSDataset) – Dataset to be fitted.
Returns: StatsTransform

transform_one(tsdata: TSDataset, inplace: bool = False) → TSDataset[source]

Transform dataset to statstransform codes.

Parameters

tsdata (TSDataset) – Dataset to be transformed.
inplace (bool) – Whether to perform the transformation inplace. default=False

Returns

TSDataset