Data Mining For Prewarning Students With Score Data

With enrollments in colleges and universities in China continues to expands, it is a major issue for universities to explore and study the characteristics of students and to make a good performance in quality control of teaching. How to improve teaching and learning is a job of management, and is also a research topic. Meanwhile, the management of college education is moving into the "standardized, information, Internet-based" direction under the impetus of the modern science and technology. Colleges and universities often have accumulated large amounts of data during teaching practice. Using data mining method to mine and analyze the large quantity of data accumulated in the implementation process of teaching, universities can get the useful information for teaching, and thus promote the quality control of teaching.In view of the current situation that tremendous impact has been brought to the quality control of teaching by the enlarging enrollment scale in universities, this thesis analyzed the data related to the early-warning students with the data mining technique, aiming to more effectively develop methods to improve the management and quality of teaching using the data analysis results.Define analysis targets, process the early-warning data, and then analyze it using statistical analysis and data mining methods to get the information of the prewarning students'department distribution, grade distribution, failed courses distribution and the changes of the prewarning students'conversion rate as well as the increasing rate. Through the classification, association rules analysis, cluster analysis of the prewarning students, we found the potential relation between the failed courses and the key factors that caused the failure, thus gained the characteristics and commonness of the prewarning students. The analysis results, providing the basis for reducing the rate of new prewarning students, increasing the conversion rate, or even predicting the possibility of appearing similar prewarning students, are helpful to the management of the prewarning students for universities.