Research And Application Of Collaborative Filtering Recommendation Algorithm Based On Neighborhood

The fast development of Internet is catching up with the busy era of e-commerce.At the same time,it also faces the problem of information overload.Film is a popular recreational and recreational project in daily life.In the face of massive film resources information of video sites,how to quickly find users' favorite movies from many movies is an urgent need to be considered.The role of movie recommendation system is to automatically recommend films to users based on their preferences and viewing behavior.Collaborative filtering recommendation algorithm is the most effective recommendation technology in the recommendation system,but there are also some problems such as data sparsity and noise data interference,which would reduce the accuracy of the recommendation.In order to solve the problems mentioned above,this paper proposes algorithm improvement on the basis of previous research.The KNN algorithm and the Slope One algorithm were optimized.The KNN algorithm based on the influence factor and the weighted Slope One algorithm based on the KNN neighborhood were proposed to improve the accuracy of the recommended algorithm and implemented in the movie recommendation system.The specific contents are as follows:(1)This paper introduces the research background and significance of the recommendation system.introduces the research status at home and abroad from the perspective of the recommendation algorithm in the recommendation system,determines the main research content of this paper.the algorithm and process of recommendation system were studied,the problems that exist in collaborative filtering recommendation algorithm were analyzed..(2)For the interference problem of traditional KNN algorithm,a KNN algorithm based on influence factor was proposed.The algorithm made the score of the closest users more weight in the final score,and we used the MovieLens data set to verify.The experimental results showed that the improved KNN algorithm not only avoid the effect of noise data on the prediction results,but also improve the accuracy of the prediction score of the KNN algorithm.(3)For users with low similarity in Slope One algorithm to weaken the effect of high similarity users,a weighted Slope One algorithm based on KNN neighborhood was proposed.First,we found the K closed neighbor users with high similarity to the target users,and then calculated the deviation between the target and other items according to the score of these K nearest neighbors,and further predicted the target user's score on the target.The experimental results showed that the improved Slope One algorithm used less but higher quality scores to help the current target users to predict their target items,not only to solve the problem of data sparsity in collaborative filtering,but also to improve the accuracy of the prediction and evaluation of the Slope One algorithm.(4)The movie recommendation system based on improved neighborhood collaborative filtering algorithm was designed and implemented.The system had the functions of publishing,deleting,modifying,querying and recommending films.It could not only view movies but also collect users' preferences for intelligent recommended movies,and fundamentally solve the drawbacks of the traditional movie information platform.