论文部分内容阅读
在考虑各种随机现象和它们的统计问题时,我们关心的往往不是其中某一项指标,而是几个其间存在着统计联系的数量指标。为了尽可能完整地搜集信息,对于每一个子样往往要测量很多项指标。主成分法就是研究如何将原始指标组合成较少的综合指标,以解释或揭示指标间的一定关系,并尽可能地表述原问题的主要信息,同时剔除多余的或依赖于其它指标变化的指标。
When considering a variety of stochastic phenomena and their statistical problems, we are often not concerned with one of the indicators but with the quantitative indicators of the existence of statistical links. In order to collect information as completely as possible, a large number of metrics are often measured for each sub-sample. The principal component method is to study how to combine the original indicators into a small number of comprehensive indicators to explain or reveal a certain relationship between the indicators and as far as possible to express the main problem of the original information at the same time remove unnecessary or dependent on other indicators of changes in indicators .