Research on Student Behavior Analysis and Prediction in Vocational Colleges Based on Multi-source Data Mining
Year: 2025
Type: conference-paper
Abstract: The student behavior prediction model based on multi-source data mining integrates feature extraction, time-series modeling and classification algorithms, and improves the behavioral evolution trend identification and risk early warning capability through the joint optimization of multi-models such as LSTM and XGBoost. Based on the sliding window and feature encoding mechanism, the model integrates multi-dimensional heterogeneous behavioral data and establishes dynamic behavioral state sequences. The experimental results show that LSTM outperforms other models in terms of accuracy, F1 value and AUC, with an accuracy of 93.5%, an F1 value of 0.936, and an AUC of 0.962, which verifies its effectiveness and stability in capturing the behavioral evolution patterns of students.
Author: Zhentao Li
Institution: Shandong Institute of Commerce & Technology
Language: English
Subfield: Computer Science Applications
Field: Computer Science
Domain: Physical Sciences
SDG: Quality education
Open Access status: hybrid
OpenAlex ID: W4417072847