Network Structure and Predictive Power of Social Media in the Bitcoin Market

36 Pages Posted: 9 Jan 2017 Last revised: 4 Jul 2018

See all articles by Peng Xie

Peng Xie

Georgia Institute of Technology - Scheller College of Business

Hailiang Chen

City University of Hong Kong - Department of Information Systems

Yu Jeffrey Hu

Georgia Institute of Technology - Scheller College of Business

Date Written: June 2018

Abstract

Following the recent discovery of social media’s predictive power in stock markets, we advance this literature by investigating whether the network structure of social media discussions can help distinguish between value-relevant information and noise. Using data from the Bitcoin market, we provide empirical evidence that less cohesive social media discussion networks are more accurate in predicting future returns. We reassure the findings with several trading simulations. Whether short selling is allowed or not, trading strategies based on sentiments weighted by discussion network cohesion yield two to three times the returns generated by random trading or strategies based on equally-weighted sentiments.

Keywords: financial technology, social media analytics, network structure, Bitcoin, prediction, topic modeling

Suggested Citation

Xie, Peng and Chen, Hailiang and Hu, Yu Jeffrey, Network Structure and Predictive Power of Social Media in the Bitcoin Market (June 2018). Georgia Tech Scheller College of Business Research Paper No. 17-5. Available at SSRN: https://ssrn.com/abstract=2894089 or http://dx.doi.org/10.2139/ssrn.2894089

Peng Xie

Georgia Institute of Technology - Scheller College of Business ( email )

800 West Peachtree St.
Atlanta, GA 30308
United States
4043696131 (Phone)

Hailiang Chen (Contact Author)

City University of Hong Kong - Department of Information Systems ( email )

Kowloon Tong
Hong Kong

HOME PAGE: http://www.cb.cityu.edu.hk/staff/hailchen/

Yu Jeffrey Hu

Georgia Institute of Technology - Scheller College of Business ( email )

800 West Peachtree St.
Atlanta, GA 30308
United States

Register to save articles to
your library

Register

Paper statistics

Downloads
568
rank
43,782
Abstract Views
1,772
PlumX