An Evaluation Index System based on Students' Behavior Characteristics based on Data Mining Technology

With the rapid development of learning analytic tools, educational data mining has emerged as a new field of research. Through the discovery of information concealed in educational data, data mining can enhance both teachers' instructional strategies and students' learning capacities. This research uses the data mining technique of Back Propagation (BP) neural network, to college students' learning behaviour data in order to forecast the outcomes of their comprehensive evaluations. According to the research, there is a strong correlation between students' learning behaviour and their overall test scores. Additionally, Decision Tree, Logistic Regression, and Naive Bayes models are created for comparison and verification. With values of MAE-0.1824, RMSE-0.3670, RAE-0.4401, as well as RRSE-0.8102, BP neural network outperforms Naive Bayes, Logistic Regression, and Decision Tree models.