Sentiment contagion analysis of interacting investors: Evidence from China's stock forum

This paper uses an agent-based model to study the sentiment contagion of interacting investors in China. We use text mining techniques to identify investor sentiment in stock forum, and then estimate the parameters of the sentiment contagion model to analyze sentiment formation from individual perspective. The empirical results suggest that the intensities of sentiment contagion in the periods of rising and falling are higher than that of the storming period. Intensities of sentiment contagion in emerging industries are higher than basic industries. A favorable macroeconomic environment will promote investors' optimism of the stock market, and high degree of investors' attention and discussions will accelerate contagion of such optimism. Once the stock market tends to be overheated, it may cause investors to worry about the market bubble, which will restrain further rise of investor sentiment. (C) 2019 Elsevier B.V. All rights reserved.