Hybrid Recommendation Algorithm Based On Content And Matrix Factorization

In the face of massive information resources,it takes a lot of time to obtain the information quickly.The recommendation system combines the user's behavior information and its own interest characteristics,it can directly recommend information that may be of interest to the user,saves time in filtering information,and solve the problem of information overload better.However,there are problems in the recommended areas such as cold start,low recommendation accuracy,and poor scalability.Traditional content-based recommendation algorithms have lower accuracy,while data sparseness and cold start problems are common in collaborative filtering recommendation algorithms.To solve the above problem,this thesis proposed a hybrid recommendation algorithm based on content and collaborative matrix factorization technique.The algorithm implements the decomposition of content and collaborative matrix in a common low-dimensional space,calculates user preference similarity and form a weight matrix while preserving the local data structure.This thesis uses a weighting technique to measure the importance of vocabulary in the text,and uses an iterative method based on multiplication update rules in parameter optimization,improved learning ability.In the target optimization function,the algorithm uses hyperparameters to controll the degree of importance of each part,and it achieves an optimal solution.Finally,the algorithm verifies and analyzes by real data set.The experimental results show that the two evaluation indexes of the proposed algorithm have different degrees of improvement,the algorithm is superior to other representative projects cold start recommendation algorithm,it effectively solves the projects cold start problem,alleviates the data sparseness and improves the efficiency of the algorithm.The paper has 19 pictures,4 tables,and 54 references.