Research And Implementation Of Online Learning Recommendation System Based On Deep Learning

In recent years,technologies such as Big Data and Artificial Intelligence have become important driving forces leading technological innovation.The deep integration of artificial intelligence and education has promoted the development of online education into a new teaching model.However,with the exponential growth and diversification of network resources,it is difficult for users to obtain the information they need in a timely and effective manner.The problem of information overload has become increasingly prominent.How to effectively solve the problem of information overload caused by massive data has become major platforms.Research hotspots.This thesis analyzes and studies the existing platforms and recommendation algorithms,proposes a hierarchical attention mechanism recommendation algorithm(DHRAA)fused with auxiliary information and designs and implements an online learning recommendation system for the existing shortcomings.The main work of this thesis is as follows:(1)The experimental data collection.In view of the lack of public data sets in the online learning field,the distributed crawler Scrapy framework is used to crawl the data of the MOOC platform of Chinese universities,and the Xpath technology is used to parse the course list data,course review data,user information,scoring information and other data to complete the data collection.(2)This This thesis proposes a deep recommendation algorithm of hierarchical attention mechanism fused with auxiliary information(DHRAA).To solve the problem of data sparsity and score prediction,this thesis proposes a deep recommendation algorithm of hierarchical attention mechanism fused with auxiliary information.Firstly,a Convolutional Neural network(CNN)is used to process the comment information of users and courses,as well as the auxiliary information such as course title,course description and teacher organization.Then,the vector representation of the comment information is extracted by using the word-level attention mechanism and the comment-level attention mechanism based on the NRPA model.After aggregating comment information and auxiliary information and extracting user implicit feature vectors and course implicit feature vectors.Attentional Factorial Machines(AFM)was used to calculate cross-combination features,and top N recommendation was completed finally according to the predicted score.The DHRAA algorithm is compared with the basic model based on the collected data set.The results show that the proposed DHRAA algorithm has better recommendation performance.(3)Designed and implemented an online learning recommendation system based on B/S framework.This thesis integrates the proposed DHRAA algorithm into the online learning platform,and uses a design pattern of separation of front and back ends to complete the development of the system based on the Django framework,and realizes the online learning module,course collection module,course recommendation module,resource download module,and coursework Module,background management module and other functions.Finally,the online learning system was tested systematically with the Hy Load automated testing tool.The test results showed that the system is practical and stable,and meets the expected goals.This thesis first improves the existing recommendation algorithm,solves the problem of data sparseness and the accuracy of score prediction to a certain extent,improves the recommendation performance,and then builds a deep learning based on the DHRAA algorithm proposed in this thesis according to user needs.The online learning recommendation system has been tested to verify the stability and practicability of the system.