An Improved Decision Tree Algorithm And Its Application On University Employment Management

Employment Report on China University Students in2012, which issued by MyCOS Institute, said that half a year after graduation,the employment rate of college students which graduated in2012only reached86%, and60%of college student which graduated in2011was not satisfied with the present job. How to improve college students’employment rate, especially Job satisfaction rate must be considered by government, educators and students themselves.The absence and development of Data Mining Techniques provides a scientific and theoretical base for government arranging employment efficiently, college reforming talents training mode around school orientation and the training goal, and college students’obtaining job fully and satisfactorily. Based on a review of the classic decision tree classify algorithm and a lot of technical literature about improved Decision Tree Algorithm, the paper presents a method of filling multi-dimensional Vacant property values by the virtue of extended Bayesian Classifying Algorithm, and put forwards an improved attribute-selection standard that makes further division on nodes while classification and replace entropy of node with the entropy of subset while having maxium information gain rate. On the basis of that, the paper builds a classifying model of employment. Then, the paper collects and disposes a massive data that connected to employment from several systems, and uses software development technology to develop an application system on employment management in which many problems about data structure and logical structure are resolved, B/S pattern and3-tiered architecture are adopted, and improved decision tree algorithm is applied. At last, the paper exacts the classification rules from decision tree and has an analysis about it. The conclusion by inference can conducts the the employment management scientifically, reasonably and effectively.