Predicting financial performance for listed companies in Thailand during the transition period: A class-based approach using logistic regression and random forest algorithm

Abstract

This study presents a class-based approach developed to evaluate the financial performance of companies that have undergone public listing on the stock market. By employing both statistical analysis and machine learning methods, the models consider two important determinants, which are the company's internal capability determinants prior to going public and the amount of funds raised through the Initial Public Offering (IPO) to predict the company’s financial performance after joining the stock market. The study demonstrates that the machine learning method (random forest algorithm) outperforms the statistical method (logistic regression) in predicting financial performance. The findings also reveal that certain determinants significantly influence the predicted financial performance in a specific period. Furthermore, the study examines the impact of IPO funds on financial performance and observes that while the first year after listing does not exhibit a significant effect, a subsequent positive correlation emerges in the subsequent two to three years, up to a certain threshold, with excessive funds potentially leading to adverse effects. Overall, the predictive models provide valuable insights for companies, enabling them to prioritize resources towards significant determinants in a specific relative year, make informed decisions, and enhance their long-term success in the stock market.

Keywords

• Financial performance
• Predictive model
• Initial Public Offering
• Internal capability
• Logistic regression
• Random forest