About this document

mRMR Feature Selection Method by toufik1986 is a document available to read on EtoBox.

Feature selection is an important problem for pattern classification. This paper studies how to select good features based on maximal statistical dependency using mutual information. It presents an equivalent criterion called minimal-redundancy-maximal-relevance (mRMR) for efficient first-order incremental feature selection. Experimental results on four datasets confirm that mRMR leads to improved feature selection and classification accuracy compared to other methods.

Author

toufik1986

Language

EN