Research On Recommendation System Based On Association Rules And Neural Network Analysis

With the rapid development of e-commerce,the recommended system has become a basic partof e-commerce sites.Recommendation system not only provides users with personalized recommendation service,but also increases the user's loyalty to the website.Recommendation system has become an important tool for competition among e-commerce websites,and recommendation system has gradually become an important research part of large-scale e-commerce websites.The core part of recommendation system is recommendation algorithm.The combination of association rules and machine learning is one of the most popular recommendation systems research methods at present.The main work of this paper is as follows:? A detailed analysis of the traditional collaborative filtering algorithm in the workflow.Respectively,demonstrate the advantages and disadvantages of user-based collaborative filtering and object-based collaborative filtering.At the same time,the application scenarios of several common recommendation algorithms are analyzed,such as singular value decomposition(SVD)recommendation algorithm,implicit semantic-based recommendation algorithm and graph-based recommendation algorithm.? The classical algorithms Apriori algorithm and FP-tree algorithm in association rules are analyzed.In this paper,the association rule mining of positive and negative items is carried out on the basis of inheriting FP-tree algorithm to scan only two times of the data sets.The experimental results show that the proposed algorithm can not only improve the time consumption,but also can automatically delete the contradictory association rules.The proposed algorithm is a supplement to the FP-tree algorithm for negative association rules.?In order to combine the user's preference model,I use the machine learning BP network fusion.The user preference model is established in detail,and the BP neural network is used to get the target user's preference model.Using the standard film score set Movieslens as the experimental data set,respectively,this paper verify the validity and robustness of the algorithm.Experimental results show that BP neural network combined with user preference model can not only improve the effectiveness of the algorithm,but also can ensure the stability of the algorithm.