Collaborative Filtering Recommendation Algorithm Based On MMTD And Users

With the rapid development of the Internet,the amount of information on the network has grown significantly.New contradictions have arisen between the production and consumption of information: it is difficult for producers to attract people's attention to the information they produce;consumers are difficult to find the information that interests them.To solve this contradiction,the recommendation algorithm came into being.The neighborhood-based recommendation algorithm is the most widely used one in the field of recommendation algorithms,which has been widely concerned and deeply studied by the academic community.As an important part of domain-based recommendation algorithms,user-based collaborative filtering algorithms have achieved a lot of research results in recent years,but these algorithms fail to make full use of positive user feedback data when measuring the similarity of users,which leads to the similarity measure between users is not accurate enough.To solve this problem,this thesis introduces the measure of medium truth degree(MMTD)and Interest bias coefficient into the collaborative filtering recommendation algorithm,which effectively utilizes the positive feedback data of users and improves the accuracy of similarity measurement between users.At the same time,when the amount of user data is large,the user-based recommendation algorithm has a relative high computation cost.To this problem,this thesis applies deep forest and Interest bias vector to user classification,which achieves the goal of lower overhead to a certain extent.The Experimental results show that the method proposed in this thesis can effectively improve the accuracy and recall rate of recommendation results,and has high practical value.The dissertation first introduces the medium mathematics and MMTD.On this basis,the user's similarity measure,the user's interest degree in items and user classification are studied deeply.The main research work and outcomes are as follows:(1)Aiming at the problem that the current user-based collaborative filtering algorithms fail to make full use of the user's positive feedback data to measure the similarity between users,a method combining user ratings and MMTD is proposed and applied to the improvement of the traditional cosine similarity and Jaccard similarity coefficient,which effectively improves the accuracy of the similarity measure between users.(2)In order to reduce the impact of subjectivity of user ratings on the performance of recommendations,an interest bias coefficient is proposed,and it is applied to the measure of user 's interest to the items with user ratings,which effectively improves the accuracy and recall rate of recommendation results,and the robustness of the algorithm.(3)Aiming at the problem of the high computation cost of collaborative filtering algorithm when the amount of user data is too large,a method combining the interest bias vector and deep forest is proposed to achieve a more accurate users classification,thereby reducing the amount of user data and effectively improving the efficiency of the algorithm.(4)Designed and implemented a movie management and recommendation system based on Java ?SSM framework and the improved recommendation algorithmin in this dissertation.This system can implement the management of user data,movie data and log data,as well as popular items,classified recommendations,guess you like three movie recommendation methods,and has high practical value.