An Analysis of College Students' Behavior Based on Positive and Negative Association Rules

Positive association rule (PAR) mining has been often used in campus data analysis, while negative association rule (NAR) mining has not, which will result in a lot of valuable information missing. This paper collects real campus data that contains the student's academic performance, e-card consumption behavior, book lending records and mental health status and analyzes them with NAR and PAR techniques. We first preprocess the data to obtain a suitable format. Then we use a method called Positive and Negative Association Rules on Correlation (PNARC) to mine NARs and PARs from these data. Finally, we obtain a lot of valuable information, for example, there is a strong negative correlation between depression and academic performance. These results are very helpful for educators to improve college students' performance.