Hybrid Recommendation System Based On Collaborative Filtering And Heuristic Association Rules

With the explosive development of computer technology at domestic and international in recent years,and the widespread use of the Internet,e-commerce has become an indispensable part of people's daily lives.The development of e-commerce is bound to be accompanied by the explosive growth of users and project.This kind of information that is flooded on the Internet that has nothing to do with user needs can make users lose their way,make users spend a lot of time searching for needs,and ultimately lead to the loss of user groups.At this time,how to quickly find items that satisfy users in the massive data becomes particularly important.Under the background of this era,personalized recommendation algorithms are emerging from time to time.By analyzing user behaviors,the potential interests of users are mined,and personalized recommendations are proactively provided to users to increase user dependence.However,the traditional personalized recommendation system also has many problems,such as data sparseness,"cold start" and low real-time performance.In order to solve the above problems,this paper proposes a hybrid recommendation system based on collaborative filtering and heuristic association rules.First,a collaborative filtering recommendation algorithm based on FCM user clustering is proposed,which combines user ratings and item features to construct a preference matrix,uses fuzzy C-means clustering algorithm to cluster users,and uses genetic algorithm to prevent convergence to local minimums.The fuzzy C-means algorithm is improved,and then the user similarity is calculated to generate a partial recommendation list.Subsequently,in view of the poor real-time performance caused by the need to scan the project database multiple times when the traditional association rule mining algorithm performs item recommendation,a heuristic association rule recommendation algorithm based on an improved genetic algorithm is proposed,which will improve the genetic algorithm and The downward closure feature of the Apriori algorithm is integrated,and individuals with better genetic genes are preferentially added to the new generation of candidate populations through crossover operations,and then the population size is further expanded through mutation operations,and another part is generated through the mined association rules The recommended results to solve the problem of low time efficiency.Finally,the recommendation results generated by the two algorithms are mixed and the TOP-N method is used to realize the final product recommendation,and the experiment verified that the hybrid recommendation system proposed in this paper can effectively improve the accuracy and reliability of the recommendation results.In summary,this article analyzes the problems of traditional collaborative filtering and association rule algorithms,and improves them.The two recommendation algorithms are combined to achieve the final recommendation through a hybrid method,which can effectively improve the data sparsity in the recommendation system and improve time efficiency...