InputTags#

classsklearn.utils.InputTags(one_d_array:bool=False,two_d_array:bool=True,three_d_array:bool=False,sparse:bool=False,categorical:bool=False,string:bool=False,dict:bool=False,positive_only:bool=False,allow_nan:bool=False,pairwise:bool=False)[source]#

Tags for the input data.

Parameters:

one_d_arraybool, default=False

Whether the input can be a 1D array.

two_d_arraybool, default=True

Whether the input can be a 2D array. Note that most commontests currently run only if this flag is set toTrue.

three_d_arraybool, default=False

Whether the input can be a 3D array.

sparsebool, default=False

Whether the input can be a sparse matrix.

categoricalbool, default=False

Whether the input can be categorical.

stringbool, default=False

Whether the input can be an array-like of strings.

dictbool, default=False

Whether the input can be a dictionary.

positive_onlybool, default=False

Whether the estimator requires positive X.

allow_nanbool, default=False

Whether the estimator supports data with missing values encoded asnp.nan.

pairwisebool, default=False

This boolean attribute indicates whether the data (X),fit and similar methods consists of pairwise measuresover samples rather than a feature representation for eachsample. It is usuallyTrue where an estimator has ametric oraffinity orkernel parameter with value‘precomputed’. Its primary purpose is to support ameta-estimator or a cross validation procedure thatextracts a sub-sample of data intended for a pairwiseestimator, where the data needs to be indexed on both axes.Specifically, this tag is used bysklearn.utils.metaestimators._safe_split to slice rows andcolumns.

Note that if setting this tag toTrue means the estimator can take onlypositive values, thepositive_only tag must reflect it and also be set toTrue.