/deker

This library is made to perform feature selection based on a method originally proposed in by Sun et al. [1]. This library specifically relates to the methodology described in [2], named DEKER for decomposed kernel regression, which includes methods for identifying optimal hyperparameter values. This library was also designed for use in the context of network inference, also described in [2], by iteratively reapplying the DEKER method for feature selection across all features of a dataset.

Primary LanguageC++OtherNOASSERTION

Stargazers