Application of machine learning in assessing financial risk of listed companies on the Vietnam stock market

Abstract

Financial risk management is essential for businesses as it helps prevent losses and maximize profits. Since this process depends heavily on data-driven decision-making, machine learning offers a promising avenue for developing innovative methods and technologies. In this paper, we compare the predictive capabilities of various machine learning models and use the LIME method to interpret how they make decisions. Data was collected from the financial statements of listed companies from 2009 to 2023. The results show that Gradient Boosting and Random Forest achieved the best performance. Additionally, LIME weights indicate that the most influential factors affecting the models' predictions are the current ratio, return on assets, debt ratio, and debt-to-equity ratio.