Application Research Of Recommended Technology In E-commerce Platform

With the rapid development of China's economy and Internet technology,e-commerce continues to grow rapidly,and the number of Internet goods grows exponentially.How to let users quickly and accurately find the products they are interested in in massive information is one of the current research hotspots.It is also an urgent problem for the current e-commerce recommendation service.Under this circumstance,the recommendation system with the core of intelligent recommendation technology emerges as the times require.The recommendation system is an important tool to solve this problem.It mainly uses some recommendation algorithms to recommend the products of interest to users.This paper mainly studies the content-based recommendation technology,Apriori recommendation algorithm and collaborative filtering recommendation algorithm,analyzes the data sparsity and algorithm complexity in the algorithm,and improves the Apriori algorithm.In the process of scanning the database,the algorithm will Elements that have been infrequently deleted are deleted in the original database,making the database scanned later smaller and smaller,reducing the algorithm runtime by reducing the number of algorithm I/Os.Combining the content-based recommendation technology with the improved Apriori recommendation algorithm makes up for the limitation of the single algorithm recommendation result and obtains better recommendation effect.In the collaborative filtering algorithm,the FCM algorithm is used to cluster users and introduce The Slope One algorithm and the implicit scoring mechanism pre-fill the scoring matrix to reduce the sparseness of the matrix;optimize the distance formula in the FCM algorithm,and replace the original Euclidean distance formula with the weighted Euclidean distance formula.The improved algorithm flow is as follows:(1)using the improved FCM algorithm to group users in the user-scoring matrix and reduce the matrix dimension;(2)using the Slope One algorithm and implicit feedback scoring to fill the scoring matrix with data.To reduce the sparsity of the matrix;(3)use the collaborative filtering algorithm to implement recommendations for the user.The improved algorithm is experimentally verified.The recommended results of the improved algorithm under different conditions are compared by using the Ali Tianchi dataset and the Audioscrobbler dataset.The improvement effect of the proposedalgorithm is verified.Use Apache Mahout's open source recommendation engine,Taste,to build a music recommendation system,and apply the recommendation ideas of this paper design to music recommendation,so that users can quickly and accurately recommend music of interest.