Critical data points-based unsupervised linear dimension reduction technology for science data

摘要

Recent advances in machine learning and data mining have led to powerful methods for the analysis and visualization of high dimensional data. This paper proposes an unsupervised linear dimension reduction algorithm named critical points preserving projection (CPPP). Selecting some key data points to represent the others has become more and more popular owing to its effectiveness and efficiency. Rather than considering all data points equally, the proposed algorithm just preserves both local neighborhood structure and global variance of critical data points. We explore a joint modification of locality preserving projection and principal component analysis to achieve these objectives. Experimental results on the UCI data sets show its good performance on pattern classification.