Research On The Application Of Data Mining In The College Admissions

As is known to all, the enrollment of colleges and universities has always been considered as the development power for their existence. According to the survey, since2008, with the decline of the number of students taking part in the college entrance examination and the expansion of college enrollment, the competition of college enrollment has become increasingly fierce and the student pool of Chinese universities has been drained. Under such situation, the recruitment of students work must use all available resources, and let these resources give full play in the recruitment of students work. However, a large number of data related to the students has been accumulated each year and they haven’t been given full use though simple statistics and analyses have been made. Therefore, to make better use of these data and improve the efficiency of enrollment, this paper tries to introduce data mining technology to college recruit students work.Data mining technology uses various methods to find implicit data patterns and information from a large number of accumulated original data. At present, the data mining technology has been very successfully applied in many fields in foreign countries, Therefore, this article applies to data mining technology and tries to explore the relationship between various factors of data and students registration based on historical admission data processing. Using decision tree algorithm through classification analysis, this article considers the student registration as the decision attributes and the students’ gender, age, area, type of the university entrance exam, score of NCEE, junior college or university and satisfaction degree to schools as the condition attributes to analyze the registration rate.Firstly, based on the brief introduction of data mining technology, the classification analysis method, C4.5decision tree algorithm and related theory knowledge, this paper gives a brief explanation of the mining software SPSS Clementine. Then, with the enrollment data of a certain university in Anhui as the example, this paper has a detailed analysis of the structure of the data and designs a scheme for a feasible sample set. Thirdly, it applies the mining software SPSS Clementine to the process of related data so as to realize the generation of sample set for mining. Fourthly, the sample set is divided into the training set and the testing set randomly for many times and use C5.0decision tree algorithm to classify the training set. So a decision tree model is constructed and the models are pruned. And we analyze the prediction accuracy and the evaluation of gain assessment as well as improvement assessment. Finally, after several repeated segmentation, classification, analysis and assessment, we select the optimal C5.0decision tree model, and export rule set. The selected model and rule set can directly reflect various factors affecting students’ recruitment. Therefore, they can provide a basis for making enrollment plans and forecast recruitment rate in the future.