WSEAS Transactions on Business and Economics
Print ISSN: 1109-9526, E-ISSN: 2224-2899
Volume 22, 2025
Early Warning Financial Distress Model for Listed Companies in Indonesia-Based Machine Learning Markov Chain Monte Carlo
Authors: , ,
Abstract: The purpose of this research is to predict the financial distress model with the Markov Chain Monte Carlo (MCMC) approach. Financial distress can have an impact on a country's macro economy. An early warning is needed to prevent financial distress by using machine learning. This research data comes from the Indonesian Stock Exchange (IDX) in 2021, with a total sample of 649 companies. To formulate machine learning, the research data is divided into two, namely, 70% for training and 30% for testing. The results of this study indicate that liquidity, sales growth, and profitability harm financial distress, and sales growth has a positive effect on financial distress. In addition, the accuracy level of determining companies experiencing financial distress in the Bayesian binary logistic model with MCMC is higher than the classic binary logistic for training data, which is 78.41% compared to 78.19%. While for testing data it is relatively the same, namely 78.35%. Since the data is unbalanced in the distress and no distress categories by less than 10%, as a result, the regression model with the MCMC procedure has the best accuracy.
Search Articles
Keywords: Early warning, Financial Distress, Liquidity, Sales Growth, Profitability, Corporate Size, Markov Chain Monte Carlo
Pages: 1370-1380
DOI: 10.37394/23207.2025.22.111