Research Of Collaborative Filtering Algorithm In E-commerce Recommendation Domain

With the rapid development of mobile Internet,the growth of the information makes people find information suitable for themselves in the vast amounts of information is more and more difficult,and the time needed to spend more and more long,which has become particularly prominent in the field of e-commerce.How to make buyers find their favorite items in a shorter time,at the same time,enables businesses to locate the potential users of their products more accurately has become a factor hindering the development of the electronic commerce,and has also became a hot spot in the study of electronic commerce.This article focus on the problem of the current e-commerce recommendation system accuracy is not high and the delay,on the basis of researching the related principle,come up with an improved algorithm,and the realization of a distributed on the Spark platform,improves the precision of the recommended,at the same time reduce the execution time of the algorithm.First,this paper analyzes the shortcomings of the most widely used collaborative filtering algorithm in recommendation system.Combined with association rule algorithm and collaborative filtering algorithm in data mining,a collaborative filtering algorithm based on association rules is implemented.Firstly,the association rules algorithm is used to classify the items with strong correlation into one category,and then the collaborative filtering algorithm is used to recommend the same category of projects.Then,on the basic of the previous analysis,a collaborative filtering algorithm based on association rules is implemented by using Python language,and the effectiveness of the algorithm is proved by relevant experimental data.At the same time,this paper studies the model of e-commerce recommendation system,Combined with the large data processing technology Spark and Hadoop,the design and implementation of distributed collaborative filtering algorithm of association rules based on Spark platform,in detail elaborated the specific experimental scheme,and through experiment verification,illustrates the association rule collaborative filtering algorithm does not reduce the performance of the related indexes because of parallelization,and time efficiency is greatly increased.The research of this paper has practical significance to improve the user's experience in e-commerce activities.At the same time,the collaborative filtering algorithm based on association rules has the same application value in other fields.Therefore,the subject studied in this paper has certain theoretical and practical significance.