Analysis method of student learning behaviour based on machine learning and data mining

With the rapid development of educational informatisation, effectively tapping its potential value to optimise teaching strategies has become an important research direction in the field of education. This article proposes a student learning behaviour analysis method based on machine learning and data mining techniques. Firstly, integrate multiple sources of data and construct a learning behaviour indicator system. Secondly, the K-means algorithm is used to group the student population and identify differentiated learning behaviour patterns; combining the random forest classification model to predict students' academic performance, and extracting key influencing factors through feature importance analysis. Additionally, introduce LSTM to explore the dynamic evolution of learning behaviour. The experimental results show that the proposed method can effectively identify high-risk student groups and inefficient learning behaviour characteristics. The research results provide data-driven decision support for teachers' precise intervention, personalised learning path recommendation, and educational resource allocation.