Analyzing QAR Data Using K-PCA

This paper analyzes the drawbacks of traditional principal component analysis (PCA) firstly, and discusses the kernel principal component analysis (KPCA) as well as its drawbacks of high complexity secondly. Then it proposes the K-PCA method. Comparing with KPCA, the method proposed in this paper could achieve dimensionality reduction with faster speed. The results show that: the proposed method performs an experiment on QAR data has a good effect of dimensionality reduction and high correct classification rate.