Research And Implementation Of Personalized Learning Resource Recommendation System Based On Big Data

With the rapid development of the Internet and information technology,the data generated by online learning resources has shown explosive growth.How to quickly dig out the learning resources that learners are interested in or need most from the massive data information,and use the resources according to learner characteristics recommending to them has practical significance for improving the learning efficiency and learning effect of learners.This paper combines the actual needs of traditional online learning platforms to design and study a personalized learning resource recommendation system based on big data.Firstly,by fully analyzing the functional requirements and non-functional requirements of learners in the personalized learning resource recommendation system,the actual needs of learning resource recommendation are investigated,the related application algorithms of learning resource recommendation,and the basic logic flow of recommendation are studied.The design is based on the overall architecture and functional module of a personalized learning resource recommendation system based on big data.Secondly,on the basis of big data analysis,a method of personalized big data learning resource recommendation combining user behavior,behavior context,user information and learning resource information is studied.In order to generate better recommendation results,this method divides learning resource recommendation into three stages: user preference acquisition,recommendation candidate set selection and recommendation result generation:(1)The purpose of obtaining user preferences is to obtain user preferences.This paper establishes a user behavior matrix,and then considers the impact of time decay and data sparsity to establish a user interest model to complete user preference acquisition.(2)Recommended candidate set screening is the idea of collaborative filtering recommendation.On the one hand,it obtains learning resources with similar user preferences from the user side,and obtains learning resources with high similarity to this project from the project side.On the other hand,it combines the k-means clustering algorithm to pass The users with high similarity are classified,and then the collaborative filtering algorithm is calculated to find learning resources similar to the user's preference.(3)To generate the recommendation result is to use the top-N recommendation idea,sort the predicted scores,and recommend the top N learning resources to the user.Then,this paper uses accuracy,recall,and F-score as evaluation indicators to design and complete comparative experiments to verify the effectiveness of the improved recommendation algorithm.Among them,Hadoop data processing services and Mahout data mining services are used to improve the overall performance of the system.Finally,using tool development Eclipse to develop a personalized learning resource recommendation system based on big data,the server side uses Java language to build basic modules,uses Java Script language and Ajax technology to achieve system interaction.