Using Nodes Atrribute Of Social Media To Predict Information

With the development of Social Media, especially the popular of Micro-blog, more and more researchers are interested in it, because of the differences between Social Media and traditional media. Especially Social Media has characters of social network, so researches of it have many important means for real world. However, many researches on using social media to predict information only take social media as the data source. They do prediction only with the number and feeling who take part in discussion. In this paper, we study nodes attributes of social media, and using them to predict information. At last, we achieved good results.The major work of this paper is presents as follows:At first, though analyzing the data of sina-microblog,it is proved some properties of social network, such as "power-law distribution" "150 rule". So we can study it through the analysis of social network.Second, we have different definitions compared with others. Definitions are no longer focus on the structure of the whole network. They are based on the predict target. Then, we make definition of static attributes and dynamic attributes. Static attributes including nodes degree, similarity of nodes and similarity based on predict target. Emotional to target and how to spreading the information about target are dynamic attributes.Third, we make a new definition:node influence based on predict target after advanced a new algorithm based on the algorithm of PageRank. With doing experiments, it proved that this definition can get a better result.At last, we selected box office prediction and stock prediction to do experiments. We find if we join nodes attributes to predict, we will have better results than ways using in current study, no matter which prediction method is used, such as gray-scale prediction, regression prediction and BP neural network prediction. The predict result is closer to actual result.As social media contains a wealth of data resources, applications and research in the broad prospects. In this study, only from an application point of view illustrates the significance of social media study, has yielded some results, of course, reflects a number of issues, which will inspire us to follow-up study. Continue to find social media research and application.