48. RECOMMENDATION FOR CHINESE MICROBLOGS

Department: Computer Science & Engineering
Faculty Advisor(s): Charles Elkan

Primary Student
Name: Chen Xie
Email: c2xie@ucsd.edu
Phone: 858-333-2354
Grad Year: 2013

Abstract
Recommendation system based on the text and social information from Microblog platform, like Twitter , Weibo etc . Applied Chinese segmentation method to transform microblogs into weighted vectors. And designed an algorithm to calculate the similarity of two users, based on their text content and user relations. Similarity calculation is based on: 1 . The number of retweets. 2 . The number of comments. 3 . The number of followers. 4 . The common relations ( topics , friends , etc ). Use TF - IDF and nomalization to transform the weighted vector. And compute the similarity based on the weighted.

« Back to Posters or Search Results