Selector

_selector

This file is a base class for the top feature selector.

class scalr.feature.selector._selector.SelectorBase[source]

Bases: object

Base class for Feature Selector from scores.

classmethod get_default_params() → dict[source]: Class method to get default params.

get_feature_list(**kwargs) → list[str] | dict[source]

A function to return top features from given scores of each feature for each class.

Parameters:: score_matrix (DataFrame) – Score of each feature across all classes [num_classes X num_features].
Returns:: List of features.
Return type:: list[str]

scalr.feature.selector._selector.build_selector(selector_config: dict) → tuple[SelectorBase, dict][source]: Builder object to get Selector, updated selector_config.

This file is an implementation of the Absolute mean feature selector strategy.

class scalr.feature.selector.abs_mean.AbsMean(k: int = 1000000.0)[source]

Class for absolute mean feature selector strategy.

It uses the absolute mean across all classes as the score of the feature.

classmethod get_default_params() → dict[source]: Class method to get default params for preprocess_config.

get_feature_list(score_matrix: DataFrame) → list[str][source]

A function to return top features using score matrix and selector strategy.

Parameters:: score_matrix (DataFrame) – Score of each feature across all classes [num_classes X num_features].
Returns:: List of top k features.
Return type:: list[str]

This file returns top K features(can be promoters or inhibitors as well) per class.

class scalr.feature.selector.classwise_abs.ClasswiseAbs(k: int = 1000000.0)[source]

Class for class-wise absolute feature selector strategy.

Classwise scorer returns a dict for each class, containing the top absolute scores of genes.

classmethod get_default_params() → dict[source]: Class method to get default params for preprocess_config.

get_feature_list(score_matrix: DataFrame)[source]

A function to return top features per class using score matrix and selector strategy.

Parameters:: score_matrix (DataFrame) – Score of each feature across all classes [num_classes X num_features].
Returns:: List of top_k features for each class.
Return type:: dict

This file returns top K promoter features per class.

class scalr.feature.selector.classwise_promoters.ClasswisePromoters(k: int = 1000000.0)[source]

Class for class-wise promoter feature selector strategy.

Classwise scorer returns a dict for each class, containing the top positive scored genes.

classmethod get_default_params() → dict[source]: Class method to get default params for preprocess_config.

get_feature_list(score_matrix: DataFrame)[source]

A function to return top features per class using score matrix and selector strategy.

Parameters:: score_matrix (DataFrame) – Score of each feature across all classes [num_classes X num_features].
Returns:: List of top k features.
Return type:: list[str]