An Important User Discovery Method For Integrating Behavior Analysis In Microblog Marketing

With the rapid development and popularization of social network,micro-blog is not only a platform for people to exchange and share information.The activities on social network platform have more and more influence on people's real life.Under this background,micro-blog marketing emerges as the times require.Its purpose is to maximize the commercial benefits of marketing activities based on social network platform.The study found that many commodity evaluation and public opinion orientation,often through a small number of people first proposed,and then by these small groups of people,these ideas spread to the whole social network.In this way,enterprises can gain their recognition and recommendation through the welfare means such as preferential or gift to a small number of people,and then spread the positive public opinion of these goods to the whole network platform,so as to make consumers have a positive impact on the purchase intention of the goods.Therefore,under the limited marketing budget,how to excavate a group of users who are conducive to brand communication becomes very important.Many scholars have studied the important users in the microblog platform.PageRank algorithm is a common algorithm to measure the importance of users.It is an algorithm to evaluate the level and importance of web pages according to the status of links between web pages.It abstracts all web pages in the Internet into small nodes,each node has internal links and external links,so the whole Internet is abstracted into a directed graph model composed of nodes and hyperlinks.The external links of each node are regarded as votes for other nodes.The higher the number of votes obtained by nodes,the higher the importance of the node,the higher the ranking should be.For example,HITS algorithm is also a common algorithm to measure the importance of users.Its ranking idea is very similar to that of PageRank algorithm mentioned in the previous section.The difference is that the evaluation of each web page by HITS algorithm involves two values: authoritative value and central value.Therefore,for the importance of each web page,the authority value andthe center value are indicators that need to be considered.However,these algorithms have some problems,such as topic drift,weight equalization of web pages and so on.The important users defined in this paper mainly include two aspects.First,important users should be interested in the target product information and pay attention to it for a long time.Second,important users themselves must have a strong influence.Then the Weibo users who satisfy the above two characteristics are the important users defined in this paper.As mentioned before,whether it is PageRank algorithm or HITS algorithm,these algorithms have problems such as subject drift,which can not measure the degree of association between Weibo users and product information,and also have weights in the process of user influence assessment.Points,ignoring user behavior and other drawbacks,can not accurately identify the important users defined in this article.In order to solve the above problems,this paper proposes an important user discovery method for the analysis of the characteristics of fusion behavior in Weibo marketing.The proposed method includes two algorithms,namely,the important user discovery algorithm based on similar topics in Weibo marketing,and the microblog marketing.An important user discovery algorithm based on impact analysis,HIDM.The IUDM algorithm first calculates the weight of the core influence factors of Weibo marketing,including: user activity and user loyalty;then,it is integrated into the PageRank calculation model,the user's probability transfer matrix is updated,and the updated probability transfer matrix is used to calculate Get the end user step by step.After the IUDM algorithm,the users who are interested in the target products and pay attention to them for a long time are screened out,and then the behavior characteristics of these users are analyzed.The original HITS algorithm is corrected by factors such as user authority,domain authority and fan contribution.Among them are users with high influence.The users thus mined are interested in the target commodity information and have long-term attention and strong influence,satisfying the definition of important users in this paper.Finally,through comparative analysis,compared with the existing mainstream methods,the proposed method can more accurately discover important users in Weibo marketing because of the more comprehensive analysis of user behavior characteristics.