Study on orthogonal tensor sparse neighborhood preserving embedding algorithm for dimension reduction

Abstract

This paper proposes the orthogonal tensor sparse neighborhood preserving embedding algorithm (OTSNPE) for dimension reduction of the high-dimensional matrix data based on the bag of visual word and in combination with the sparse representation. OTSNPE applies sparse coding to local characteristic quantification of data through completion of within-class sparse representation and preserves the supervised local geometrical information effectively. Finally, the experimental result of the real high-dimensional matrix data set verifies the effectiveness of the algorithm.