Research On Key Technology Of Product Recommendation Base On Orange-Collar

Social network and online shopping are more and more closely related to people’s life.Online shopping in the proportion of everyday retail is also increasing.In order to increase the network sales,business began to use blogs,microblogs and other ways to promote their products in social network.More and more people tend to obtain information from the network,especially from the social network.All these resulted in the production of social purchasing guide.There are a group of people who are called orange-collar in the groups of social purchasing guide.They publish blogs about shopping in the social network to guide their fans online shopping,and they can gain profits from the business.But because of shorter development time and the different growth pattern of each orange-collar,how to make better recommendation for products is needed to be solved quickly,and that is also the important problem in social purchasing guide.So this paper study the key technology of product recommendation base on orange-collar,and propose three method to solve the problem of orange-collar positioning,how to recommend better orange–collar for final users and how to recommend better orange–collar for business.The main contributions and innovations of this research include:First,this thesis proposes a method named orange-collar positioning.This method not only could recognize orange-collar and no-orange-collar,but also could make it clear that the classification of the product that orange-collar recommends.It could posit orange-collar.Base on this method we could study on orange-collar recommendation.Secondly,this thesis proposes a method that named orange-collar recommendation for final users.This method could solve the problem that how to find the most suitable orange-collar for persons.The method has taken two hands in consider.The one hand is whether orange-collar matches person’s requirement,the other hand is orange-collar‘s influence in SNS.This method could provide a solution for consumers to get better shopping experience.Thirdly,this thesis propose a method named orange-collar recommendation for merchant.Based on the item based collaborative filtering algorithm and content-based recommendation algorithm,this thesis analysis the products that to be recommended and the products recommended by orange-collar,get the similarity of them.Then taking the orange-collar’s influence in consider.At last getting the suitable recommendation.This method could provide a solution for merchants to get better goods promotion.Finally this thesis has done experiments to verifier the effectiveness of the methods.The experiments base on the messages from orange-collar persons,sub-orange-collar persons,normal persons in Sina Weibo.