Corporate Failure Risk Classification Based on Multilayer Perceptron

The pandemic and political instability lead to heavy impact on economic and social environment for many countries. This also brings negative impacts to corporates. Failure prediction contributes to useful information for decision making process of managers, financial institutions and shareholders and give the early warning in predicting bankruptcy risk of enterprises. Machine learning have been recognized as efficient methods in classification and prediction comparing to the traditional statistic models. In this paper, we consider the performance of machine learning models in failure forecasting. We compare Multilayer Perceptron to Support Vector Machine and Logistic Regression in risk prediction with Vietnamese scenario. We use 24 financial ratios that are taken from the annual financial report of Vietnamese companies during time 2017–2019 for risk forecasting. The outcome shows the high accuracy of machine learning methods and the outperforming of Multilayer Perceptron comparing to Support Vector Machine and Logistic Regression. It contributes to great information channel for decision makers to give some implications in risk protecting.