论文部分内容阅读
三、判别分析(Discriminant Analysis) 在实践中常常遇到判别问题,例如:根据造林地的若干特征判别其立地类型;根据症状和化验指标,确定病人患何种疾病;根据气象资料预报未来某一时期的天气等等。与聚类分析不同,解决这类问题需要掌握历史资料,对已有的样本作出明确的分类,即事先取得关于类的知识,然后对一个新样本(待判样本)的类属作出判断,因此判别分析常被称为有监督的分类。
Discriminant Analysis Discriminant analysis often encounters the problem of discrimination in practice, for example, identifying the type of site based on several characteristics of the site for afforestation, determining the patient’s disease according to the symptoms and laboratory indexes, and predicting a certain future disease based on the meteorological data Weather and so on. Different from cluster analysis, it is necessary to master the history data to solve such problems, make a clear classification of the existing samples, that is, obtain the knowledge about the classes in advance, and then judge the generic type of a new sample (pending sample) Discriminant analysis is often referred to as supervised classification.