A supervised class-preserving Laplacian eigenmaps for dimensionality reduction

A supervised class-preserving Laplacian eigenmaps (SCPLE) algorithm is proposed. In this algorithm, two neighbor graphs, intra-class graph and inter-class graph, are constructed, whose edge weights determined by class label information and adaptive thresholds. By maximizing the weighted neighbor distances between heterogeneous samples and minimizing the weighted neighbor distances between homogeneous samples, this algorithm maps homogeneous samples closer and heterogeneous samples farther in the low dimensional space. Experiments demonstrate the superiority of the proposed algorithm against the classical LE, DVE and S-LE algorithms.