基于信息增益比的决策树用于毕业生就业预测  

Research of Applying the Method of Decision Tree Based on Information Gain Ratio to College Student's Employment Forecasting

作　　者：程昌品[1] 陈强[1] 

机构地区：[1]广东教育学院计算机科学系,广东广州510310

出　　处：《计算机仿真》2010年第2期299-302,共4页Computer Simulation

基　　金：广东省科技计划项目(2007B010200038)

摘　　要：关于毕业生就业的预测研究,对毕业生就业及学校招生与教育教学管理工作具有指导作用。针对目前毕业生就业预测存在的不可靠性,提出了一种基于信息增益比的决策树应用于毕业生就业预测分析的方法,通过计算样本空间的信息熵和测试属性的期望信息熵以求得测试属性的信息增益比,从而构造出基于信息增益比的决策树,据此能够准确预测毕业生的就业情况。以广东教育学院的毕业生就业情况为例进行预测,实验结果表明所用方法的有效性和可靠性。Graduate employment projections play a guid role in the employment of graduates, school enrolment and education management. Because the current employment rate of graduates is not reliable, a decision tree based on the ratio of information gain is proposed to forecast employment of graduates. By calculating the entropy of space samples and testing the entropy of attribute the test attribute the decision tree is achieved, which can accurately fore- cast the employment situation of graduates. The method is used in Guangdong Institute of Education graduate employment forecast, and the results proved the effectiveness and reliability.

关 键 词：信息增益比 决策树 就业预测 有效性 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]