A Study Upon The Influence Of A Single Micro-Blog Based On Multilayer Structure

Micro-blog was introduced into China in 2009. Most of the studies had focused on the macro level—users, meanwhile the research on the micro level—single blog has been poorly studied. It varies in the repost quantity of different micro-blogs released by different micro-bloggers, and different micro-blogs of the same micro-blogger. Therefore, how to help people to find out the micro-blogs and micro-bloggers with great influence and the key factors that affect repost quantity is worth researching.Based on the above, combining with the multi-level structure of micro-blog, the study on the micro-blogger’s influence can be divided into several stages. First is to analyze the key factors affecting repost quantity, and to devise mathematical models. Then, based on user’s attributes, combining the attributes of a single micro-blog, build a Hierarchical Liner Model of the micro-blog. Finally, according to the characteristics of micro-blog transmission, build a predicting model. The study optimized the effecting factors by using the H index and took repost quantity as the objective measurement of micro-blog’s influence.This paper analyzed the data of Sina micro-blog and the main conclusions are as follows:(1) the indexes provided in the paper are more reasonable in measuring the influence of micro-blogger. (2) resorting to the Hierarchical Liner Model, it can explain the difference better than a simple linear model. (3) the significant factors affecting influence of a single micro-blog can be obtained by continually optimizing the model. (4) the predicting model can fit well the actual data, predict well the trend of the influence of a single micro-blog. The rationality of the decrement and the probabilities of a retweet are validated.The main innovation points in the paper are as follows:(1) built a new index system. To optimize the indexes upon micro-bloggers, we introduced H-index and built a new index system bearing macro and micro structures. (2) built models of the micro-blogger’s influence to quantify the influence and a Hierarchical Liner Model of the micro-blog to explain the relations between influence and indexes.