Collaborative Filtering Recommendation Algorithm Research Based On Model

With the advent of the "Internet+",information data is rapidly accumulating,and the amount of data is exploding.How to get the required information quickly and efficiently has become an urgent problem in the rapid development of the Internet.The existing information retrieval and filtering technology cannot fully meet the needs of all kinds of users.How to provide personalized information services has become a research hotspot in the development of the Internet.Personalized recommendation techniques analyze user preferences based on user historical behavior.It provides personalized information services to meet user needs and gives users a better service experience.It has become one of the important tools for users to obtain information.This paper focuses on the analysis of the model-based collaborative filtering recommendation algorithm.The main innovations and work of this paper are as follows:(1)An explainable matrix factorization algorithm based on neighbor relationship is proposed.The traditional matrix factorization model has good accuracy,but there are some shortcomings that the model cannot explain.In this paper,explainability parameters are introduced in matrix factorization,explainability constraints are added,and neighbor relationships are added in latent space.Those methods solve the problem of explainability of matrix factorization model and improve the explainability of recommendation algorithm.Firstly,we consider how to generate an explainability matrix more perfectly and reasonably,and introduce different weights of similar users when generating the explainability matrix.Then,we consider the influence of neighborhood relationship in the latent space,introduce neighborhood relationship into the latent space,and make the similar neighbors in the latent space closer to each other in the distance.Finally,the neighbor weights are further introduced into the latent space vector to achieve better accuracy and explainability.(2)A neural network collaborative filtering model with basic side information is proposed.The neural network has good fitting characteristics,can approximate any continuous function,and effectively capture the relationship between users and items.Based on the user rating matrix,this paper uses neural network to mine the key interaction relationship between users and items-collaborative filtering relationship.On this basis,the basic information of users and items is directly introduced,and the collaborative filtering information and basic information are integrated into the neural network as a whole.Further considering the statistical relationship existing in the basic information,the extracted statistical information is introduced into the neural network to strengthen the relationship between the user and the item,so as to achieve a better recommendation performance.