In this study, the CSI 300 Index in China mainland and original articles from authoritative stock WeChat public accounts are investigated regarding their relations. First, a sentence-level sentiment classification approach for analyzing investor sentiment polarities in text corpus is proposed by expanding synonyms. Then, the Granger causality test is utilized to examine the impact of sentiment index on the stock price and volume-values. It shows that the influence of overall investor sentiment on volume-values is more rapid than that on stock price and the impact of positive sentiment is found to be more lasting than the negative in both stock price and volume-values. Furthermore, it is worth noting that there is a dual-stage phenomenon in the impact of positive sentiment on volume-values, which indicates that some investors react to positive information immediately while others may choose to wait and follow the trend.
Zhao, Haiyuan; Wang, Dannuo; Wang, Mingyan; He, Xinrui; and Jin, Jian, "Mining the Impact of Investor Sentiment on Stock Market from WeChat" (2019). WHICEB 2019 Proceedings. 61.