Application and Research of Data Mining Based on Improved PCA Method

The LAMOST (large sky area multi-object fiber spectroscopic telescope) is one of the national key scientific projects. It will yield 10,000~20,000 spectra per observation night. Automatic spectral analysis and recognition focused on helping astronomers finding their interesting celestial objects. become desirable and necessary. In this paper an efficient data mining application based on improved Principal Component Analysis (PCA) is proposed, which has less computational complexity. Massive spectral data are clustered after dimensionality reduction by PCA. The singular spectra candidate then can be found out and identified by template.