Abstract: In order to analyze the learning quality of undergraduate students in communication engineering major in Beijing University of Posts and Telecommunications (BUPT) and improve the training program, a score analysis method based on improved clustering and association rules mining is presented, and data mining methods are used to analyze the correlation among the students′ performance in different courses Taking the scores of classes of 2009~2011 undergraduates in communication engineering major in BUPT as the sample data, the proposed mining method is used to obtain the network of association rules so as to analyze the connection among courses, to determine the key courses and to evaluate the effectiveness of teaching Meanwhile, taking the National College Entrance Examination (NCEE) scores of class 2011 as a sample, association rules between NCEE scores and university scores are mined in order to analyze students′ performance after the entrance The results can provide some insights for the improvement of teaching methods, which is quite important to improve students′ learning quality

Key words: score analysis, data mining, clustering, association rules