基于R软件rpart包的分类与回归树应用

The Application of the Classification and Regression Tree Based on the Package rpart in R-Lamguage

对于许多分类和回归问题,二叉树(Binary Tree)提供了有趣而又形象化的方式来研究数据,它主要是按照一定的规则拆分自变量,而完成对因变量的合理分类,进一步可以对未知分类进行预测。在主要介绍递归分割(Recursive Partitioning)和回归树(Regression Tree)在R软件中应用的同时,对一前列腺癌数据使用生存分析和分类与回归树相结合的方法做出分析,并得到了对于疾病诊断和预防较有指导意义的结论。

For many problems ooncerning classification and regression, the method of ＂binary tree＂ has provided an interesting visualization approach in research. The basic idea of such methods is to split the response variable according to certain rules, thus we can get a reasonable classification of the response variable, and make prediction to the unknown classifications based on new samples. This paper mainly introduces the implementation of recursive partitioning and regression tree in the package rpart of R language, then makes an analysis to a medical data using classification and regression tree and survival analysis, and finally gets some useful instructions on the diagnosis and prevention of illness.