Integration of KPCA and parallel coordinates for visualizing classification

Combined with pattern recognition and information visualization technology, this paper proposes a visual classification method based on KPCA and parallel coordinate plot KPPCP. This method maps the raw data space into high-dimensional feature space by means of nuclear function, then the feature space is deal with PCA, finally the processed data is visualized in parallel coordinates. The experiment show that it can effectively extract the non-linear features from the raw data , enlarge the differences between the various categories, provide Interactive visualization, enhance the understanding of experts on the classification process and participation so as to get more effective classification.