This paper presents the first work in the task of author profiling for Vietnamese blogs. This task is important in threat identification and marketing intelligence. We have developed a Vietnamese Blog Profiling framework to automatically predict age, gender, geographic origin and occupation of weblogs' authors purely based on language use. The experiments on the blogs corpus we collected show very promising results with accuracy of around 80% across all traits. © 2009 IEEE.

Pham D.D., Tran G.B., Pham S.B.

223.pdf    Gửi cho bạn bè

Từ khóa : Geographic origins; Threat identification; Weblogs; Blogs; Internet; Linguistics