改进决策树算法在高校就业管理中的应用研究

发布时间：2018-06-12 00:31

本文选题：C4.5算法 + 贝叶斯分类　；参考：《华东理工大学》2013年硕士论文

【摘要】：麦可思研究院(MyCOS Institute)发布的《2012年中国大学生就业报告》显示,2012届大学生毕业半年后的就业率仅约为86%,而在已就业的大学生中,2011届有60%的大学生认为目前的工作与自己的职业期待不吻合。如何促进大学生就业,提高大学生就业率,尤其是提高大学生就业满意率,是政府部门、教育工作者以及学生本人都应该思考的问题。数据挖掘技术的出现与发展,为管理决策部门科学有效地部署就业工作,为高校围绕学校定位、围绕人才培养目标改革人才培养模式,为大学生充分就业、满意就业,提供了科学的理论依据。本文在简要回顾几种经典的决策树分类算法,大量查阅有关分类改进算法文献的基础上,提出延伸贝叶斯定理实现一次性填充多维空缺属性记录,提出分类过程中进一步划分深层结点,选取具有最大信息增益率情况下的子集熵代替深层结点的信息熵,据此创建大学生就业分类模型。然后,本文从多个系统收集并整理了大量与就业相关的数据,并基于这些数据利用软件开发技术,解决了软件开发中有关数据结构和逻辑结构的难点,采用B/S模式和三层架构开发模式,开发出了围绕上述改进算法的就业管理应用系统。最后,本文利用系统抽取出了就业分类规则,对规则进行了分析与解释,所获得的相关结论可以科学地、有效地、及时地指导就业管理工作。
[Abstract]:The Employment report of Chinese University students 2012 released by MyCOS Institute of MyCOS Research Institute shows that the employment rate of Chinese college students in 2012 is only about 866.In 2011, 60% of the students who have been employed think that their current employment is similar to that of their counterparts. Your career expectations don't match. How to promote the employment of college students, improve the employment rate of college students, especially the employment satisfaction rate of college students, is a problem that should be considered by government departments, educators and students themselves. The emergence and development of data mining technology, It provides a scientific and theoretical basis for the management decision-making department to arrange the employment work scientifically and effectively, to reform the talent training mode around the orientation of the school, to reform the talent training mode around the goal of talent training, and to provide the scientific theoretical basis for the full employment and satisfactory employment of the college students. In this paper, based on a brief review of several classical decision tree classification algorithms and a large number of literature on the improved classification algorithms, an extended Bayesian theorem is proposed to realize the one-off filling of multi-dimensional vacant attribute records. In the process of classification, the paper further divides the deep nodes, selects the subset entropy with the maximum information gain rate to replace the information entropy of the deep nodes, and then establishes the employment classification model for college students. Then, this paper collects and arranges a large number of employment-related data from many systems, and based on these data, using software development technology, solves the difficulties of data structure and logical structure in software development. An employment management application system based on the above improved algorithm is developed by using the B / S model and the three tier structure development mode. Finally, this paper extracts the employment classification rules systematically, analyzes and explains the rules, and the relevant conclusions can guide the employment management work scientifically, effectively and in time.
【学位授予单位】：华东理工大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：G647;TP311.13

【参考文献】