C4.5算法在民族地区高校毕业生追踪调查中的应用研究——以河池学院为例  

Study on the Application of C4.5 Algorithm in Follow-up Investigation of University Graduates in Ethnic Areas——Examples of Hechi University

作　　者：梁达友[1,2] 董荣胜[1] 王泓刚 

机构地区：[1]桂林电子科技大学,广西桂林540001 [2]河池学院,广西宜州546300

出　　处：《科技创新导报》2014年第10期252-256,共5页Science and Technology Innovation Herald

摘　　要：部分民族地区高校毕业生就业是个难点问题。本文介绍了ID3、C4.5和CART三种决策树算法,并选取河池学院毕业生追踪调查数据,分别建立ID3、C4.5和CART分类器进行预测比较。结果表明,C4.5预测分类器具有最小计算量和最高准确率。通过测试,C4.5分类器预测分类准确度和精度与实测数据相符,能够较好地满足应用要求。C4.5分类器分类规则表明,道德素养、专业素质和实践能力是用人单位录用毕业生的主要标准。预测结果同时也为学校创新人才培养模式,提高毕业生就业能力提供一定的决策支持。Employment of university graduates is a difficult problem in some ethnic areas. Based on the introduction of three kinds of decision tree algorithm, including ID3, C4.5 and CART, and selection of the data of Hechi University graduates as follow-up survey data, this paper had compared with the predictions in the establishment of ID3, C4.5 and CART classifier. The results showed that C4.5 prediction classifier has the characteristics of the smallest amount of calculation and maximum accuracy. By testing, C4.5 classifier’s accuracy and precision in the aspect of prediction classification corresponded with the real data, which can better meet the requirements of application. C4.5 classification rules showed that moral quality, professionalism and practical ability are the main criteria for employers hiring graduates. Forecast results also provide a decision support for innovative talents’ cultivation model and employability improvement of graduates for schools.

关 键 词：教育数据挖掘 追踪调查 预测分类 C 4 5算法 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]