Research On Recommendation Algorithms Integrating Trusted User Feedback

With the development of mobile Internet technology and the rise of social networks,the Internet has become the main way for people to obtain and publish information.While explosively increasing information has brought convenience to people's lives,overloaded information has also begun to haunt people's lives.For this reason,personalized recommendation incorporating social information has become not only a research hotspot in the field of machine learning and data mining,but also an intelligent service tool widely used by the Internet and social networking platforms.The existing personalized recommendation methods incorporating social information are mainly based on matrix factorization-based collaborative filtering recommendations as prototypes,incorporating social information reflecting user preferences into matrix score prediction,which greatly improves the accuracy of the recommendation system.However,these existing recommendation methods rely too much on explicit feedback such as user ratings,ignoring feedback information that is scarce but directly reflects users' likes and dislikes.To this end,based on the existing probability matrix decomposition model that fuses social information,this article will design a new objective function and recommendation algorithm to use these feedback information reflecting users' likes and dislikes to reduce the user's disliked items in the recommendation list Probability.The work of this paper is described as follows:(1)A recommendation algorithm based on user feedback is proposed.This method uses recommendation based on probability matrix factorization as the basic model.It introduces these sparse feedback information that accurately reflects users' likes and dislikes into the user score prediction function,and improves the sparseness of the implicit feedback information through social network trust calculation.Optimized the recommendation list without reducing the overall performance of the algorithm,and realized effective filtering that made users dislike or dislike the content.Experimental results on real datasets verify the effectiveness of the algorithm.(2)A recommendation algorithm based on trusted users is proposed.Although the user's likes and dislikes can greatly reduce the probability that the user's objectionable content appears in the recommendation list,the vast majority of users do not have such direct feedback information that reflects their likes and dislikes.To this end,this article will rely on trust computing to generate associations between users through attention,forwarding,etc.,and disseminate the likes and dislikes of a few users to these associated users,in order to make full use of these scarce but effective feedback information to reduce associated user recommendations.The probability of objectionable results in the content.Experimental results on real datasets verify the effectiveness of the algorithm.(3)Based on the above algorithms,this paper designs and implements a recommendation algorithm application system that integrates trust user feedback.Based on the processing of user ratings and feedback,this system uses the recommendation algorithm proposed in this article to effectively implement the user's favorite movie list recommendation,especially when the user inputs feedback such as "not interested",which is related to it The content is effectively filtered.