Research on Public Opinion Effecting on Stock Price During Crises Based on Model Checking

Recent studies have shown that news and investors’ comments on social media against listed companies significantly impact stock price movements. Listed companies suffer abnormal stock price movements and tremendous economic losses in public opinion crises. Most existing studies combine news and investors’ sentiments with technical indicators to predict stock price movements to help investors optimize investments. However, it is urgent for listed companies to predict stock price movements during crises and understand the impact of multiple factors on stock price movements during crises, thereby using effective response strategies to stabilize stock prices. This study proposes a method to predict stock price movements during crises based on model checking. We integrate the public opinion factors by considering the interaction among investors, media and listed companies as crisis managers based on the Situational Crisis Communication Theory. The explicable rules of stock price movements are extracted by random forest algorithm from objective data of previous crises and formalized as Computation Tree Logic formulas (φ). The database is modeled into a verifiable formal model (M) and formalized as a Kripke structure. The model checker NuSMV is used to verify the rules in actual situations to predict stock price movements and provide early warning automatically. The proposed method achieves superior performance with ACC of 77.53% and 78.43% in the victim and preventable crises, respectively. The response strategies of listed companies significantly impact stock price movements during crises in the Chinese stock market. The explicable rules of stock price movements provide decision support to develop proper crisis responses for listed companies during crises.