Abstract: To establish the system of graduate fellowships evaluation of graduate education in universities is one of the most important problems mechanism reform. Through data analysis to the graduate fellowships evaluation information related properties, this paper constructs the scholarships evaluation system data sets, and uses the information gain classification algorithm of C4.5 decision tree data mining. Finding some interesting modes of the graduate student grants by the specific case study, it can promote the establishment of effective graduate student funding mechanism for providing the scientific decision support of the training mechanism reform.

Key words: decision tree, data mining, training mechanism reform, graduate fellowships evaluation

摘要： 研究生奖助学金评价体系的建立是我国高校研究生培养机制改革中的重要问题。通过对研究生奖助学金评选信息相关属性的数据分析，构建了研究生奖学金评价体系数据集，并采用基于信息增益的C4.5决策树分类算法进行数据挖掘。通过具体实例研究发现了研究生奖助学金评选中的一些有趣模式，对建立有效的研究生资助机制，为培养机制改革提供科学的决策支持起到了一定的促进作用。

关键词: 决策树分类算法, 数据挖掘, 培养机制改革, 奖学金体系评价