Research On Demand Effect Of Product Network And User Network

The recommendation system plays an increasingly important role in assisting users in various decision-making processes,such as product recommendation,music audition or news reading,on Amazon.com,Taobao.com,YouTube,Netflix,Yahoo,JD.com,Last.fm and IMDB,etc.It plays an important role in popular Internet sites.Most e-commerce sites offer a variety of forms of recommendatio n,from simply displaying the most popular items to using sophisticated data mining techniques to provide useful information.Due to the existence of various recommendation systems,a large number of products are not isolated.In these websites,the products are nodes,and the recommended links between products can be regarded as the side of the product network.In addition,as social platforms such as Facebook,Weibo,and WeChat build a wide range of user-to-user links,product evaluation information is efficiently disseminated and shared in the link network.This article focuses on product networks and user networks,focusing on "similarity" from e-commerce platforms,online video and music platforms for three areas of research.First,previous studies have not analyzed the impact of similar product networks.We explored the similarity network by mining product recommendation links from Taobao,increasing the understanding of the commercial value of the similarity recommendation system.By analyzing the characteristics of similarity product networks,based on the complex network perspective,we analyzed the demand effects of network attributes such as the degree of product analysis and network center potential,and quantified the commercial value of similarity links.The research results indicate the location in the product network.The importance of this,and we observe that the degree of product penetration network has a positive demand effect and the central potential has a negative demand effect.Our research expands the research on product networks by focusing on "similarity" and analyzing different types of product networks.In the context of similar product networks,based on text mining,we quantified the spillover effects of UGC and MGC on products,and found that the similarity of product reviews and the similarity of product descriptions will have a negative impact on product demand.The demand shows a stronger impact than UGC.Second,previous studies lacked an analysis of the homogenous effects of the reviewer population in online video platforms.Based on the homogeneity theory,we comprehensively consider three similarities,user preferences,comment content and comment emotions,and construct a user preference similarity network.Through the weighted dynamic network method,the results show that the reviewer's similarity network requires video.Influential,that is,the homogeneity of the commentator group has a negative impact on video demand,and the analysis finds that the diversity of the comment content has a positive impact on video demand.In addition,few studies have focused on the relationship between emotional similarity and demand diffusion in online social video commentary.We analyzed user interactions,the effects of similarity emotions contained in each other's comments,and found negative and positive comments.More diverse and differentiated emotions will increase the viewing needs of the video.Third,we conduct research from a similar product network to consider a user network,and then combine the two networks.Unlike previous co-buying networks,we have improved the accuracy of forecasts through product-like networks and user network information.Compared with the prediction of the benchmark,the demand information and network attributes such as Pagerank,degree centrality and local clustering of similar music neighbor nodes placed in the prediction increase the user network related information,such as the number of fans,the degree of user penetration,and the improvement.The predicted effect of product demand.It has obvious commercial significance for online e-commerce platform or music platform to improve operational efficiency.