Research And Design Of Multi-Objective Optimization Resource Recommendation System For Adaptive Learning

With the rapid development of information technology and continuous innovation of educational ideas,smart education has developed into an important field of educational research.In smart education,adaptive learning is a commonly used solution to address students’ personalized learning needs.In order to solve the core problem of adaptive teaching,which is how to accurately recommend learning resources that are most suitable for each student’s learning needs and level,AI-based recommendation algorithms are widely applied in adaptive learning resource recommendation,helping students better obtain personalized learning resources and services.Based on the characteristics of university course teaching,this thesis studies the design of a test question recommendation system that can be used for adaptive learning,taking into account students’ personalized learning basis and the characteristics of test questions such as knowledge points,ability dimension and difficulty.This thesis considers several objectives,such as knowledge point coverage,knowledge continuity,learning ability growth,learning initiative protection and stimulation.The main content includes the following two aspects:(1)A Parallel Test Recommendation Algorithm based on Multiple Decision Trees and Cognitive Diagnosis with DSFM-based LSTM Prediction Network(MDT&CD DSFM-LSTM),which combines a DSFM-LSTM(Deep Structured Feature Model based on Long-Short Term Memory)prediction network,has been designed to address the problems of inaccurate recommendations and insufficient personalization in learning resource recommendation.The algorithm integrates the recommendation results of multiple decision trees and cognitive diagnosis,and predicts students’ states using a prediction network,in order to recommend the most suitable practice questions to students.Experimental results show that the recommendation results and evaluation effects of the MDT&CD DSFM-LSTM algorithm are superior to those of the commonly used improved decision tree algorithm.It can help students better understand knowledge points,improve learning ability,and has broad application prospects.(2)The requirement analysis and function design related to learning resource recommendation in intelligent evaluation system are designed and implemented.To achieve accurate positioning of the recommendation system,a test question bank labeling method with 36 attribute fields and corresponding data structures were designed.Considering the efficiency and stability of the system operation,this thesis adopts Web technology and a front-end and back-end separation B/S architecture.Based on the My SQL database,the test question library and database for the learning resource recommendation system were built,and adaptive recommendation business functions such as test question management,paper management,and test paper publishing were implemented.The improvement of these system functions can not only help better manage test questions and publish tests but also perform data analysis,automatically discover students’ problems and deficiencies,and provide targeted exercise question recommendations,realizing intelligent counseling for course learning,improving the quality and effectiveness of students’ effective learning.