Relative Local Mean Classifier with Optimized Decision Rule

摘要

Local mean classifier can achieve good effect for many real problems and need not explicitly determine the prototypes beforehand. However, it still can not be comparable with human being in classification on the noisy, the sparse, and the high dimensional data. This paper proposes an new approach, called relative local mean classifier(RLMC), to overcome this problem by utilizing the perceptual relativity. It finds k nearest neighbors for the query sample from each class and then performs the relative transformation over all these nearest neighbors to build the relative space. Subsequently, each local center is computed in the relative space, which is then applied to perform the classification. Experimental results on both real and simulated data validate the proposed approach.