Predicting Herding Behavior in Retail Investors using a Random Forest Algorithm

Investment success hinges not solely on market outperformance, but on the presence of a well-defined financial strategy and disciplined behavior. This integration of unbiased patterns guides investors toward their desired financial destinations. The investor’s decision-making process is inherently associative, with choices influenced by an array of possibilities often tainted by irrational behaviors. Among these behavioral biases, herding bias looms large, significantly impacting the investment decision-making process. This research aims to predict herding bias by employing the Random Forest algorithm and delving into the factors that contribute to its prevalence. An online questionnaire was meticulously constructed and distributed to 500 individual investors residing in urban areas, with 318 responses forming the basis of our analysis. Our findings reveal that a lack of self-efficacy is a common trend among investors, prompting them to follow the investment choices of others. The Random Forest algorithm exhibits remarkable predictive accuracy, registering an impressive 98.44% in forecasting the herding bias among retail investors in urban settings.