Based On Microblogging Media Platform Of A Microblogging Popularity Prediction And Related Microblogging Found

Sina micro-blog has played an important role in the dissemination of information in our daily life, the daily number of active users of Sina micro-blog has reached 66 million 600 thousand by the end of 2014. The sina micro-blog Media Platform has a number of fans, their influence is big, the news for the supervision becomes very difficult.How to find the popular micro-blog from the amounts of micro-blog is very significant. It can help people to understand the social dynamics in real time, so that some government agencies can make a guide about some negative emotions in advance, and deal with the negative emotions which cannot be broadcast in advance. It will make the society more stable. According to the trend of the development of micro Bo, the scholars pay more and more attentions to the micro Bo hot topic detection in recent year, micro blog public opinion analysis are hot research.This thesis mainly constructed a microblogging Media Platform real-time monitoring system, which is used to predict the hot topic of microblogging in the future. This thesis mainly on Sina micro-blog news public number of data acquisition, and the data analysis, through this analysis we construct prediction model. By calculating each micro Bo forwarding, reviews the growth trend to predict the tweet in the current hot degree. And by the clustering algorithm,we can find the related tweet.The main contributions of this paper are as follows:1. Through Sina micro-blog API interface, we collect real-time microblogging data and collect the data of Sina micro-blog every ten minutes by Sina micro-blog API interface.and then analyze Sina micro-blog, This thesis found the factors that affect the trend of microblogging, through this analysis, This thesis found that microblogging comments, forward growth number has a concave curve downward trend along with the passage of time.2. A predictive model is structured, the model consists of the long-term trend microblogging, trend cycle, the influence, microblogging comments before ten minutes and forwarding. The model can calculate a micro Bo number of comments and forwarding number in the trend of the future, then this thesis give the popular degree formula, through the formula we compute the popularity of a micro Bo.3. K-means algorithm was improved, the algorithm set by entering a cluster center and a threshold to find the distance clustering center within the threshold of text. By this algorithm we can find and popular microblogging related micro Bo.