Research On A Social Recommendation Algorithm That Integrates Multiple Relationships

With the advent of the era of big data,the information on the Internet has grown rapidly.Although this gives people more choices,it also brings about information overload.The recommendation system is generated under this background.Social recommendation algorithm based on social network is a popular method in recommendation system at present.The existing social recommendation algorithms only consider the impact of one relationship on the recommendation results.However,there are many relationships among users in real social networks,and each relationship has different effects on recommendation.Therefore,the introduction of a social relationship in recommendation will inevitably affect the accuracy of recommendation results.In order to explore the impact of multiple social relationships between users on the recommendation effect,in this thesis,a social recommendation algorithm that integrates multiple relationships research is carried out.And experiments and analysis on the real data set is conducted.The specific research content and research results are as follows:1.The mass diffusion algorithm is a kind of collaborative filtering algorithm based on the neighborhood.In this thesis,a mass diffusion algorithm based on multi-subnet composite complex network model is proposed.A multi-subnet composite complex network model is used to construct user product rating network and user social network.And through the loading operation of the composite network,the network is transformed into a space vector.The interrelationships between nodes are mapped into multidimensional vectors in space vectors.In the end these networks were merged into a new composite network.Through the principle of mass diffusion,the initial energy of the goods purchased by the target user is spread on the newly synthesized composite network.According to the final energy obtained by the product after the spread,the Top-N product is recommended to the user.The experimental results show that the introduction of research on multiple social relationships between users can effectively improve the accuracy of the recommendation algorithm,and the social relationship can more significantly improve the recommendation accuracy of inactive users who purchase fewer products.2.Matrix factorization algorithm is a typical representative of model-based collaborative filtering recommendation algorithm.A matrix factorization social recommendation algorithm based on multiple social relationships is proposed.In this thesis,multiple social relationships between users are introduced in the traditional matrix decomposition process.The user product rating matrix is decomposed into two matrices,namely the user feature matrix and the product feature matrix.Considering that the social relationship between users will affect the recommendation effect,the user feature matrix is reorganized according to the user's social relationship matrix.The target parameters containing the restructured user feature matrix are calculated.The gradient descent algorithm is used to find the minimum value of the loss.The optimal user feature matrix and product feature matrix are obtained.The product prediction score matrix of the user is obtained through the product of the two.Finally,the target user is recommended based on the predicted score.The experimental results show that the introduction of research on multiple social relationships between users can effectively improve the accuracy of the recommendation algorithm.Moreover,the social relationship can more significantly improve the recommendation accuracy rate of inactive users who purchase fewer products.