Semi-Supervised Local Discriminant Analysis

Aiming at the disadvantage of unsupervised method and supervised method,a linear dimensionality reduction method called Semi-supervised Local Discriminant Analysis(SLDA) is proposed.When there is no sufficient training sample,local structure is generally more important than global structure.SLDA utilizes the labeled data points to infer the local discriminant structure,as well as the intrinsic geometrical structure inferred from both labeled and unlabeled data points at each local area.Experimental results on ORL and Yale face recognition demonstrate the effectiveness of the algorithm.