Research On Collaborative Filtering Recommendation Method With Category Information

The rapid development of the Internet makes people get rich and colorful content information.At the same time,it is becoming more and more difficult to get information to meet the needs of Internet users from the Internet which full of mass information,the problem of information overload is becoming more and more serious.Although the search engine has greatly alleviated the overload problem,it still can't meet the individual needs of the people.The recommendation system comes into being,which is designed to help people find the information they are interested in.It provides different personalized services for different users.The traditional recommendation system is to explore the relationship between the user and the project,which can provide users with the items of interest.The collaborative filtering recommendation system is different from the previous recommendation technology,which not need to rely on the feature extraction of recommended objects to determine the user interest,and can solve the problem of user interest transfer.The recommendation is highly personalized recommendation,so it has become the most widely used recommendation technology.Collaborative filtering is the most successful and widely used information technologies which makes personalized predictions by exploiting the historical behaviors of users.The accuracy of the recommendation depends on effectiveness of the similarity measure.The methods of traditional similarity measure,which mainly concern with the similarity of the common ratings but ignore the category information in the rated items,are suffering from data sparsity problem.To address this issue,in this paper,we study the collaborative filtering recommendation algorithm combing category information and user interests.The specific work is as follows:(1)According to the popular bias problem,this paper makes full use of the item category information,propose a new approach of interest similarity measurement based on classification information,this approach is combined with the traditional user similarity measure to obtain an improved method of user similarity measurement.The improved user similarity measure is more practical and more accurate.(2)In order to alleviate the data sparsity problem,improve the accuracy of finding nearest neighbors,this paper puts forward a method for filling user-item matrix based on category information,which makes up the deficiency of the traditional matrix filling method.(3)We compared the collaborative filtering recommendation method with category information with existing collaborative filtering methods,and the simulation experiments are carried out on the MovieLens data set.The experimental results demonstrate that,compared with other recommendation algorithm,the algorithm wo proposed can make up for the deficiency of the traditional method and reduce the influence of data sparsity problem,improve the accuracy,diversity and novelty.