Xiaozhong Liu

Title: Connect Language, Network and Culture Bubbles: Cross-Twitter and Weibo Information Recommendation and Community Comparison

Date: Nov 14, 2014

Time: 12:30pm – 1:30pm

Room: Wells Library Rm LI030

Abstract: All the popular social media systems, i.e., Facebook, Twitter and Weibo, can only reach a limited group of users. For instance, Twitter is forbidden in China because of political reason, and Weibo is not being used in US and Europe because of language barrier. In this project, we design innovative algorithms, models and system to efficiently compare Twitter and Weibo communities from different perspective, which can be used to better understand the similarity and difference of very large user communities. For instance, given the same topic “Iraq War” or “iPhone 6”, Twitter and Weibo users may access, mention or discuss it in very different way. In this study, we use the sophisticated text mining and graph mining algorithms along with big data indexation methods to compare Twitter and Weibo users. The results can be very useful for social scientists to better understand the large user communities. Meanwhile, we proposed a new method to efficiently integrate Twitter and Weibo data, and generate a Pseudo Global Social Media Network (PGSMN), which, to the best of our knowledge, is the first global social media network. By using PGSMN, we can implement global information recommendation, i.e., recommend Twitter information/topic to Weibo (Chinese) users and vice versa.