Classroom Learning Behavior Analysis and Grade Prediction Based on Data Mining

Under the background of big data era, with the rapid development of information technology and the large-scale application of artificial intelligence and cloud computing and other technologies, all walks of life have accumulated massive data, which often contains valuable “knowledge” and “information”. The development of educational informatization and online education has produced super-large data. It has become a practical need of colleges and universities to use educational data mining technology to dig out valuable information and present it to learners and teachers, so as to improve students’ academic performance and teaching model. In this study, education data mining technology is used to collect students’ demographic characteristics, personal characteristics, learning environment and learning input information from various information systems, such as school educational administration management system, student management system and student personal information system, so as to construct a student academic performance prediction framework. Then by using Bayesian network, decision tree, support vector machine and neural network four algorithms to establish analysis models respectively, analyze the factors that affect academic performance. Research has found that among the four aspects that affect students’ academic performance, learning engagement is the most important. The results of the four models show that the more students devote to study, the better their academic performance. The experimental data of this study came from a local university. Whether this conclusion is universal to other universities in China needs further research.