The Optimization Of Movie Recommendation Based On Collaborative Filtering Algorithm

With the development of the information technology,the increasing amount of information in the daily information search process has brought great inconvenience to users who want to find information,and the problem of information overload has become increasingly prominent.Therefore,searching for information which is useful for users in complex information data and recommending it has become a topic of great concern in the content service industry.And film recommendation is a hot and difficulty topic in current research.Watching movie is a popular entertainment method.It is difficult for movie viewers and movie content service platforms to choose and to recommend movies to viewers.Based on this point,from the perspective of the film service content platform,this paper studies how to find movies which match user's interest in the massive movie resources.Based on the review of the related research on recommendation problems,this paper analyzes the advantages and limitations of the existing collaborative filtering recommendation algorithm,and proposes several methods to improve the collaborative filtering algorithm.The main research contents include:(1)A collaborative filtering algorithm based on CBR(Case Based Reasoning)is proposed for the problem that the collaborative filtering algorithm cannot be recommended without marks by searching similar cases from case library.Then the scores of similar cases are assigned to the target case,when the target case is filled,and the recommendation is finally realized.Finally,simulation experiments were carried out using the Movielens dataset widely used in the film recommendation study.The experimental results show that the improved CBR collaborative filtering algorithm can still be recommended without the browsing record,and it is more accurate than other algorithms.(2)To improve recommendation accuracy,a recommendation method which uses user trust is proposed.By incorporating user trust into traditional collaborative filtering algorithms to improve recommendation accuracy.Different from other methods,the algorithm no longer finds all the paths between users to determine the degree of trust,but only uses the optimal trust path between users as the basis for determining the trust degree to improve the recommendation efficiency.The experimental results based on dataset show that this method can effectively improve the recommendation accuracy of collaborative filtering algorithm by 5%.(3)For the problem that the algorithm recommendation efficiency is not high,an improved k-means algorithm is used to deal the data.The users are grouped before recommendation in this way,and it can reduce the scope of finding similar users and improving recommendation efficiency.The improvement of the k-means algorithm is that k initial points can be automatically selected to avoid the clustering result falling into the local optimal solution.The experimental results show that the proposed algorithm is more efficient than the traditional collaborative filtering algorithm,and the recommended speed is increased by about eight times,while the accuracy is only reduced by 1.5%.