【作者】 宋向红；

【导师】 石磊；

【作者基本信息】 郑州大学 ， 计算机应用技术， 2011， 硕士

【摘要】 随着数据挖掘技术的成熟及其应用领域的扩展,不少普通院校的研究人员已开始将其应用于普通高校的管理中并得到了相关研究结论。由于成人高校和普通院校相比在学生来源、教学模式、管理方式等方面有所不同,直接将普通院校的研究结果应用于成人高校中有所不妥,所以采用数据挖掘技术对成人高校的生源情况、学生的评教记录以及学生的考试成绩等进行研究具有重要意义。本文首先采用决策树分类算法ID3对平顶山教育学院往年的生源情况进行分析,生成分类规则,得到结论：年龄较小且收入较低或一般的教师是学院成人教育生源的主体。研究表明,增加新专业,扩大生源范围势在必行。然后采用关联规则的Apriori算法对教师评教数据进行挖掘,产生了相应的强关联规则,结果表明成人高校的学生相对较为成熟,经验丰富的老教师、职称较高的副教授和知识丰富的具有硕士学位的教师,相对评价较好。最后采用聚类分析的k均值算法对考试成绩进行聚类,得到了簇中心和个类数。结果显示：如果优秀率、良好率、中等率、及格率和不及格率符合正态分布,说明教学效果良好,学生对课程内容掌握较好。更多还原

【Abstract】 With the perfection of data mining techniques, it has been applied in many fields. The researchers in some common colleges applied it to the college management and got some valuable conclusions. Because of the difference of students’source, teaching modes, management methods and so on, the research results got from the common colleges can’t be applied to the adult colleges directly. So, it is significant to research the students’source conditions, teaching evaluation records, examination results and so forth by means of data mining.Firstly, by means of decision tree classification algorithm ID3, the students’ source condition in former years of Pingdingshan Institute of Education was analyzed, and the classification rules were generated. Research results indicate that the main student source of adult education colleges is the teachers who are young and having lower or middling income, so it is necessary for the adult education colleges to add new specialties and extend student source fields.Secondly, based on the Apriori algorithm of association rule, the teachers’ teaching evaluation data was mined, and the strong association rules were got. Research results indicate that the students of the adult colleges are relatively mature, and the experienced teachers, the teachers with associate professor professional title and the teachers with master degree and abundant knowledge have good teaching effect.Lastly, the K average algorithm of cluster analysis was applied to cluster to the examination results, and the cluster centers and the number of cases in each cluster were obtained. Research results indicate that if the rates of excellence, good, middle, pass, and fail is subjected to normal distribution, it shows that the teaching effect is good and the students master the course well.更多还原