Detecting and Analyzing Influenza Epidemics with Social Media in China

In recent years, social media has become important and omnipresent for social network and information sharing. Researchers and scientists have begun to mine social media data to predict varieties of social, economic, health and entertainment related real-world phenomena. In this paper, we exhibit how social media data can be used to detect and analyze real-world phenomena with several data mining techniques. Specifically, we use posts from TencentWeibo to detect influenza and analyze influenza trends. We build a support vector machine (SVM) based classifier to classify influenza posts. In addition, we use association rule mining to extract strongly associated features as additional features of posts to overcome the limitation of 140 words for posts. We also use sentimental analysis to classify the reposts without feature and uncommented reposts. The experimental results show that by combining those techniques, we can improve the precision and recall by at least ten percent. Finally, we analyze the spatial and temporal patterns for positive influenza posts and tell when and where influenza epidemic is more likely to occur.