QuantileOutliersTransformer

class pyoptex.analysis.transformers.quantile_outlier_transformer.QuantileOutliersTransformer(factors=(), Y2X=<function identityY2X>, random_effects=(), threshold=1, stat='norm')

Transformer using the Quantile-quantile plot method to detect outliers based on a threshold from the ideal value. Drops the terms one-by-one based on the largest deviation as long as the value is above the threshold.

The fit_transform function fits the data and removes the detected outliers. During regular transform, nothing happens as this should only remove training outliers.

Note

It is extended by OutlierTransformerMixin.

Attributes

thresholdfloat

The threshold for dropping terms on the deviation from the quantile line.

statstr

The distribution to use for the quantile-quantile plot.

errors_np.ndarray(1d)

The errors (pred - y) for a simple model fit.

outliers_np.ndarray(1d)

A boolean array marking which rows are considered outliers in the training dataset.

__init__(factors=(), Y2X=<function identityY2X>, random_effects=(), threshold=1, stat='norm')

Creates the outlier transformer

Parameters

factorslist(Factor)

A list of factors to be used during fitting. It contains the categorical encoding, continuous normalization, etc.

Y2Xfunc(Y)

The function to transform a design matrix Y to a model matrix X.

random_effectslist(str)

The names of any random effect columns. Every random effect is interpreted as a string column and encoded using effect encoding.

thresholdfloat

The threshold for dropping terms on the deviation from the quantile line.

statstr

The distribution to use for the quantile-quantile plot.

Methods

QuantileOutliersTransformer.fit(X, y)	Fits the data.
QuantileOutliersTransformer.fit_transform(X, y)	Fit the transformer to the data and apply the transformation.
QuantileOutliersTransformer.preprocess_fit(X, y)	Preprocesses before fitting the data.
QuantileOutliersTransformer.quantile_outliers(...)	Quantile-quantile outlier detector based on the desired distribution and a threshold value.
QuantileOutliersTransformer.set_output(*[, ...])	Set output container.
QuantileOutliersTransformer.transform(X, y)	Ignore any transformation as the outlier detection only applies during training.

Attributes

QuantileOutliersTransformer.is_fitted

Checks whether the regressor has been fitted.