Abstract: Highlights•The joint information granule considers more possibilities.•The neighborhood information entropy measures are constructed.•A nonmonotonic algorithm is designed.•Information Gain is introduced to perform preliminary dimensionality reduction.
Loading