A Personalized Movie Recommendation System Based On Collaborative Filtering

In recent years,the rapid development of Internet technology has provided people with an increasingly wide range of services,greatly enriching their cultural and recreational life,but the rapid growth of data and information has caused the problem of information overload.In the information explosion of the Internet era,how to provide valuable information to users quickly and effectively has become an important issue that cannot be ignored.At the same time,with the development of the economy,the needs of users are increasing and becoming more and more personalised and diverse.In this context,personalised recommendation system technology has emerged and is growing rapidly.The collaborative filtering algorithm is one of the most widely used and successful recommendation system technologies,but it also has some disadvantages that cannot be ignored,such as sparse data,cold start,low recommendation efficiency,and long time consuming.Therefore,this paper presents an in-depth study and improvement of the collaborative filtering algorithm in order to provide better personalised recommendation services.This paper first shows several of the most commonly used algorithms in recommendation systems and highlights their advantages and disadvantages.Then,this paper proposes a new recommendation algorithm based on the traditional collaborative filtering algorithm with improvements,and completes a personalised movie recommendation system based on the proposed algorithm,with the main work including:1.A hybrid recommendation algorithm based on user-based collaborative filtering and item-based collaborative filtering.The single recommendation algorithm has disadvantages,for example,the user-based collaborative filtering suffers from user cold start and data sparsity,while the item-based collaborative filtering algorithm suffers from item cold start and the novelty of the recommended items is poor.In order to solve the problems of single recommendation algorithm and improve the quality of recommendation,this paper proposes a hybrid recommendation algorithm,the main work includes:(1)using the modified cosine similarity formula to improve the accuracy of similarity calculation in collaborative filtering algorithm;(2)weighting fusion of user-based collaborative filtering and item-based collaborative filtering algorithm,and then adjusting the weighting coefficient to get the best recommendation effect.Experiments demonstrate that the hybrid algorithm proposed in this paper improves the system recommendation performance.2.The traditional collaborative filtering algorithm suffers from serious data sparsity problem and low accuracy of user similarity calculation.This paper improves it and proposes a collaborative filtering algorithm based on user clustering and improved user similarity.To address the data sparsity problem,this paper adopts the K-means++aggregation technique,using the K-means++ algorithm to cluster users and recreate the user-item rating matrix within each cluster,and finally using the reduced-scale matrix to implement the collaborative filtering algorithm;secondly,to address the problem that popular objects in the recommendation system affect the calculation of user similarity Secondly,to address the problem that popular objects in the recommendation system can affect the accuracy of user similarity calculation,this paper introduces the penalty factor of popular products in the calculation of user similarity to reduce the influence of popular products,thus improving the recommendation performance of the algorithm.3.This paper develops a personalized movie recommendation system based on the hybrid recommendation algorithm.Firstly,the feasibility and functional requirements of the system are analysed and discussed;secondly,the workflow of the system is elaborated from the overall architecture and database construction,and the recommendation service process is introduced in detail;finally,the system functions are implemented and tested,and the test results show that the system runs smoothly and can provide effective personalised movie recommendation services for users.