Research on Student Behavior Pattern Mining and Management System Based on Deep Learning

According to the practice of students' work, the paper makes a statistical analysis of students' achievement, attendance, comprehensive quality and borrowing books. The kettle (ETL) is used to extract, purify, and transform the data, and it is loaded into the oracle database for storage and management, and various rule bases and algorithms are used to parse and the data. By using graphics, tables and other visual tools, an intelligent analysis platform for students' work under the environment of big data is established. The data structure and statistical method are used to express the user's behavior characteristics, obtain the specific behavior information, and build the behavior database according to the specific time-granulated way. This project proposes a clustering mining method based on behavioral characteristics based on behavioral data sets. The research results show that the visualization platform of college students' behavior analysis built with big data can realize accurate diagnosis of students' learning situation, personalized learning analysis, intelligent analysis and decision making, and early warning of disciplinary violations. It greatly improves the teaching quality of the school and achieves the purpose of “teaching students according to their aptitude”.