Research And Implementation Of Movie Recommendation System Based Nearest Neighbor And Weighted Slope One Algorithm

In the 21 st century,the rapid development of the Internet has allowed countless people to feel the fun brought by the Internet.While surfing the Internet,people also suffer from the annoyance brought by the Internet.Internet information is becoming more and more complicated.Tens of billions of new messages are updated every day.This makes people unprepared to search for the information which they want.Despite the rapid development of search engines and search technology,users are still faced with difficulties in sorting and using various types of information.In this context,the recommendation technology has become more and more important and come into being.In e-commerce sites for sale,e-commerce recommendation systems are used in the sale of goods,replacing traditional salespeople in business.In addition to e-commerce sites,people watching the Internet and watching videos and movies through the Internet have become the main forms of amateur leisure.The major video media companies have risen suddenly,such as Iqiyi,Douban,etc.,in order to win a place in the fierce competition.This sites efforts to improve the user experience.As we all know,showing consumers movies that are in line with their interests is one of the important measures to increase user experience and increase users' stickiness.Therefore,all major video sites are striving to do a good job with the user's good interaction,put the recommendation technology into the background development module,to create a better user satisfaction,and then obtain profit.This theme aimed to develop an online movie recommendation site,analyzed and designed and completed the relevant function modules of the online movie site,and ensured the applicability of the Slope one algorithm on the basis of guaranteeing all basic functions of an online movie site.To make it more suitable for movie recommendation sites.Firstly,the similarity between users was calculated by cosine similarity,the nearest neighbor of the target user was selected,and then the user weight was adjusted according to the user's evaluation quantity,and the weightcoefficient was introduced.The users involved in the calculation were replaced with the nearest neighbors and combined with the Slope one algorithm.Experiments verified that the improved algorithm has a little improvement in precision and reduction of time-consuming.Then the improved algorithm was placed in the recommendation module of the sites back-end,which improved the service level of the entire movie recommendation sites so that the consumer can use it more conveniently and satisfactorily.