Design and evaluation of intelligent student management system based on multi-modal data

With the advancement of information technology, the demand for intelligent management systems in the field of education is becoming more and more urgent. Based on the urgent need for an intelligent management system in the field of education, this study puts forward the hypothesis that an intelligent student management system integrating multimodal data can effectively improve the efficiency of education management and promote the personalised development of students. In terms of methodology, deep learning algorithms are used to innovatively integrate multimodal data such as students' academic performance, behavioral performance, and psychological state, and a comprehensive evaluation model for students is constructed, using Python as the main programming language, using TensorFlow framework to build and train deep learning models, and using MySQL databases for data storage and management. The results showed that the system performed well in data fusion, with an accuracy rate of 95 %, which was significantly higher than the 80 % accuracy of the traditional system. In the personalised recommendation function, the accuracy rate is increased to 90 %. Comparative experiments have found that after using the system, students have an average improvement in academic performance of 12 % and a reduction in behavioural problems of about 15 %. This fully verifies the effectiveness of multimodal data fusion in the intelligent student management system and provides a solid practical basis and direction for the intelligent development of education management.