Perceptual Relativitybased L ocal Mean Center Classifier

Xianfa Cai, Guihua Wen, Jia Wei, Zhiwen Yu, Yongming Cai


1Despite k -nearest neighbors and its variants perform very well in many applications, they usually suffer from such limitations as sensitivity to noisy, sparse and imbalance data which results to a dramatic performance degradation. However human cognition has its unique ability to deal with these issues from the different perspective. Motivated by this character, this paper proposes a feasible strategy called Perceptual Relativity-based Local Mean Center Classifier by using the relative transformation to local mean center classifier (RLMC). Firstly, relative transformation will be performed over the training samples to build the relative space and find k nearest neighbors in the relative space. The advantage of relative transformation is that it improves the distinguishing ability among data points and diminishes the impact of noise on classification. Experimental results on both real and simulated data suggest that the proposed approach often gives the better results in classification and robustness.


Full Text:



  • There are currently no refbacks.