摘要
主成份是多元统计分析中的一个重要概念，通过样本主成份与原样本指标的相关分析，从理论上探讨了主成份分析法中样本信息的损失量。
Principal component (PC) is a main concerpt of multi statistic analysis. The PC is usually choosed as following: Omitting the PC whose chracteristic root nears zero, or omitting the later principal components by the given valve value r , when the characteristic roots of the former PC add up to r . Both the above methods lose some sample information which will be discussed in the second part, and the distribution of the loss information between samples will be discussed too. Practical example is given in the third part.
出处
《湖北大学学报（自然科学版）》
CAS
1996年第2期138-141,共4页
Journal of Hubei University：Natural Science
关键词
特征值
样本信息
主成份分析法
损失问题
Characteristic root
Principal component
Correlative coefficience
Sample Information
