Repost Number Prediction of Micro-blog on Sina Weibo Using Time Series Fitting and Regression Analysis

Sina Weibo, as the most popular micro-blog platform in China, has become a major source of network hot events and sensitive public opinion. This paper presents a scheme to predict the repost number of micro-blog message. Curve fitting and time-series model are used for the prediction. In order to improve the predicting precision, an empirical correction model are built by utilizing the prediction data of 3200 micro-blog messages using least square and second-order polynomial regression methods, which takes the daily periodic fluctuation of reposting probability into consideration. By experimental verification, the proposed scheme can predict the repost number of micro-blog message accurately.