Research On Personalized Hybrid Recommendation Algorithm Based On Preference Feature

Since the twenty-first century,the explosive growth of data,so that people's lives are full of information resources.However,it is difficult to extract valid information from these messages.How to quickly expand the resources of fast and accurate information for users to meet the individual needs of users,has gradually become a number of researchers and network users concerned about the hot issues.The recommendation system is a new technology that can quickly find effective information and accurately discover user interests,and also gives users a good user experience.In this paper,I have analyzed some of the commonly used algorithms in the personalized recommendation system.By analyzing the preferences of project attributes and the weight of attribute preference,it can be seen that there are obvious differences in the preference of project attributes for different users.Based on the statistical analysis of the project attribute preference,the user 's preference for project attributes can be determined.Then,a collaborative filtering recommendation algorithm(UPIACF)based on the user' s attribute preference is proposed according to the project attribute preference feature.On the basis of the research of this algorithm,I compare the experimental results and compare the results of this algorithm with the traditional UCF algorithm.The results show that the UPIACF algorithm has better improvement than the traditional UCF algorithm in the recommended accuracy rate,recall rate and accuracy.Although the UPIACF algorithm improves the accuracy of the proposed results in a certain proportion,the algorithm itself only considers the user's score data and the preference characteristics of the project.For example,for the user has watched,has been rated film,its project attributes include love,comedy,science fiction,etc.,but did not consider the user's preferences attribute attributes,such as the user's age,gender,hobbies,occupation and so on.Therefore,in order to further improve the recommended accuracy,recall rate,etc.,in the study of this paper from the user preferences and project preferences and other characteristics of multiple complementary perspective,ready to proceed with complementary features of the user and project preferences feature multiple attributes To the personalized hybrid recommendation algorithm,I intend to mix the recommendation algorithm based on user attributes(UAA)and the collaborative filtering algorithm(UPIACF)based on the user's preference for project feature attributes to form a personalized hybrid recommendation algorithm UPIACF_UAA.In this paper,we study the hybrid algorithm not only on the hybrid algorithm but also on the hybrid algorithm.Finally,the experimental results of the hybrid algorithm proposed in this paper are compared with the algorithm based on user preference feature attribute(UAA)and the traditional user-based cooperative filtering algorithm(UCF),The final experimental results show that the improved hybrid method proposed in this paper can get more ideal results than the UCF_UAA hybrid algorithm.