Design And Implementation Of A Video Recommendation System Based On Linear Regression Learning

With the advent of the Internet age,online video learning is being welcomed by more and more people because it is not limited by time and location.This learning method has become a part of modern education and teaching system.Online video learning resources are becoming increasingly huge,and the massive learning resources in the network bring users selection obstacles.How to choose the video they want is a difficult problem for learners.First of all,although there are many learning resources on the network platform,they are not managed systematically,so learners are easy to get lost in the face of information overload.It has become very difficult to quickly find the videos that they are interested in learning in the video learning website platform.Secondly,network resource learning does not reflect the core idea of people-oriented,and there is a certain conflict between resource overload and learners' personalized needs.With the wide application of collaborative filtering recommendation technology in recommendation system,the above problems have been well solved.However,the reduction of recommendation accuracy caused by scalability and data sparsity has become a problem of conventional collaborative filtering technology.Based on the research of linear regression and collaborative filtering algorithm,this paper develops a learning video hybrid recommendation system.In this paper,we first build a recommendation algorithm model based on improved collaborative filtering,then analyze the requirements of the recommendation system,design the overall architecture of the system,specific modules and database,and divide the system into user module and background administrator system module.Finally,we use Hadoop,HDFS,Map Reduce and other technologies to achieve user upload learning video,like,comment,score,query,etc Search and attention function,as well as the administrator of the video management function.After testing,the whole system runs normally,which improves the accuracy of the user's prediction score,and the effect is in line with the expectation.