Design And Implementation Of E-Commerce Recommendation System Based On Diversified Information Fusion

In recent years,with the rapid development of Internet technology and Web service,network information has spread rapidly and increased sharply.Although it is convenient for human life,there is also the problem of"information overload",so the recommendation system technology came into being.With the continuous growth of e-commerce platform users,the recommendation system is also gradually improving and making great achievement in the field of e-commerce.The main purpose of the recommendation system in the e-commerce platform is to increase the click rate and order transaction rate by pushing items that meet the user preferences.Scholars at home and abroad have carried out many researches on recommendation algorithms,but there are still problems such as sparse data and inability to dynamically track user interests,which can easily lead to decreased user satisfaction and poor experience.This thesis relies on the traditional collaborative filtering algorithm to develop a recommendation algorithm of diversified information fusion to solve the above problems.The main work is as follows:1)For the data sparse problem of the user rating matrix,we fully consider the rating habits of different users and the influence of the item attributes on the rating data.The missing value of the matrix is filled by calculating the user preferences weight value for each item attribute and the average rating of the rated items by different users.2)When calculating the similarity of users,we introduce the popular item inhibitor into the Pearson correlation coefficient to deeply mine the personal interests and hidden preferences of users,and effectively reduce the impact of popular items on the accuracy of the recommendation algorithm.At the same time,the time factor weight is introduced according to the human forgetting curve function to dynamically track the changing trend of user interests,and the two similarity calculation results are effectively weighted and combined.3)Based on the Spring Boot service framework and related technology stacks,we design and implement an ecommerce recommendation system,and apply the recommendation algorithm of diversified information fusion to the e-commerce system to realize personalized recommendation service and effectively improve the satisfaction of system platform users.