The Analysis of Student Experimental Behavior Based on Multi-source Information Fusion.

At present, with the rapid development of Internet technology and the arrival of the big data, more and more smart labs are put into use. The analysis of multi-source data in the labs has a wider application scenario. Experimental teaching based on big data platform has become more and more popular. Learning and using the multi-source massive data generated by the experimental teaching platform has become a hot issue in the field of educational research and application. Based on the surveillance video of the circuit laboratory of the smart labs, this paper uses the open-source posture recognition framework OpenPose to process the human posture in the video, and then obtain the human bone point data. It trains the appropriate classifier to classify and recognize the behavior based on machine learning, and analyzes the correlation between the behavior data and the performance data to establish a learning prediction model, which on the one hand provides scientific guidance for student management, On the other hand, it provides valuable decision-making information for experimental teaching reform.