The Application Of Similar Product Set In Economy System

With the fast development of Internet, we are now in the era of overwhelmedinformation. How to precisely offer information that users are interested in is obviouslyan essential research topic. In current E-commerce System, PersonalizationRecommendation is widely used to recommend product to users. The binary relationbetween users and product information are built in Recommendation System, and user’spotential for certain products is obtained by data mining of current choice-makingprocess and similarity relationship, then personalized recommendation can be achieved.Detailed analysis, regarding the application of Recommendation Algorithm inE--commence System, showed two features: massive objectives (products and users)and sparse data of user behavior. Therefore, two main problems are demonstrated: theusability of recommendation algorithm and recommendation model can hardly beassured; the Matthew Effect occurs, resulting in profound concentration of beneficiaries.Based on text mining and product database, this article proposes a productclustering algorithm，which is called Similar product set algorithm. This algorithmapplies text mining and behavioral targeting technology to converge similar products.By analyzing the whole products, similar product set are built between the product layerand category layer. Proper change to the fineness of similar product set is necessary iftarget and need changes. Except for solving the sparseness of user behavior database,this algorithm also can provide solutions to price analysis, pricing and diagnosis ofrepeat products.This article is divided into four main parts:1. Analysis about the problem, sparseness of user behavior data, and solution alreadyexisted in recommendation algorithm of E-commerce.2. Similar product set algorithm based on text mining is proposed to solve this mainproblem above.3. Data analysis and experiment are carried out to check the proposal. 4. Applicable solutions associated with similar product set are given to problems likeprice analysis, pricing and diagnosis of repeat products. And finally, come toconclusions after experiment.