Weibot

February 2018

Overview

Weibot is a social media robot on Chinese Weibo platform. Weibo, the twwiter like social media platform, literally means micro blog, has more than 165 milllions of activated user (per day) and the most popular user and famous actress Xie Na has more than 100 millions of followers (according to weibo user statistical report by September 2017).

Instead of a social bot with designed behaviors, this robot is intended to response to its autonomy through learning. A nature language processing method bases on Markov Chain was considered; this bot wil learn from existed history post of one human user and create its own post (prediction from estimated Markov chain model).

In order to build the learning model(markov chain), history post of one selected user needs to be acquired. With web scraping technique, this robot can scrape all recorded post of any given id, take Xie na as example, all of her posts in past 9 years has been scraped and store here. Since the learning model part is still under development, all post has been analyzed and result in a word frequency map for now.

Future Development

Keep working on the learning model till it can get some reasonable post meanwhile take Xie Na's new post as cross references. Xie, as the No.1 weibo user keeps more than 5 to 10 posts everyday. And the NLP markov chain library considered for this project is Markovify