Machine Learning Models for Predicting Financial Distress

Abstract

Difficulties in business liquidity and the consequent financial distress are usually an extremely costly and disruptive event. For this reason, this study attempts to provide a set of features that can help us predict the sustainability of a company. This study involves the building of a financial prediction system which after training on a set of companies' historical final accounts (ranging over a period of 3 to 5 years), the models built are then capable of evaluating the nature of another companies' financial data. Consequently, the company's financial position in the following financial period is predicted (whether a company is active or failing). After predicting firm financial health, the outputs of the Decision Tree, the Naïve Bayes classifier and the Artificial Neural Net are evaluated to see which algorithm is the most accurate in finding a set of features for this problem. The research findings over a reallife datasets confirmed the strength and ability of the proposed model in predicting eminent business failure. Moreover, Base-year and year-over-year comparison both produce good results, therefore both techniques can be used for financial analysis. The optimal feature set included ratios from all categories, meaning, company profitability, liquidity, leverage, management efficiency, industry type and company size are all crucial to distress prediction. The prototype implemented in this study attempts to answer open questions, such as whether ML techniques are capable of predicting financial distress and whether financial ratios and industry variables are indicative of financial sustainability.

Supervised Learning for the Prediction of Firm Dynamics

“…The latter stream of literature (financial distress prediction)-pioneered by Fantazzini and Figini [36]-deals with the problem of predicting default probabilities (DPs) [77,12] or financial constraint scores [66]. Even if these streams of literature approach the issue of firms' viability from slightly different perspectives, they train their models on dependent variables that range from firms' bankruptcy (see all the "bankruptcy" papers in Table 4) to firms' insolvency [12], default [36,14,77], liquidation [17], dissolvency [12] and financial constraint [71,92].…”

Unveiling the diverse efficacy of artificial neural networks and logistic regression: A comparative analysis in predicting financial distress

“…We also note that the accuracy has improved compared to the training phase, and this is considered a suitable. However, the model showed great weakness in identifying financial distress cases, as it failed to identify (3) cases, and it was able to identify (5). On the other hand, it was able to identify cases of non-distress very well, by determining (15) cases of non-distress, and it failed to determine one case.…”

“…Initially, statistical methods were employed by early researchers in the field of bankruptcy prediction. Despite the continued utilization of statistical methods, certain scholars have embraced the application of Machine Learning techniques, including neural networks, to forecast corporate failures (Bonello et al, 2018). The onset of the second phase, which commenced in the late 1980s, marked a pivotal moment wherein numerous scholars embarked on investigations aimed at ascertaining the efficacy of non-parametric methodologies in prognosticating bankruptcy risk.…”

Crunching Numbers, Making Decisions: Artificial Intelligence and Statistics for Financial Distress Forecasting in Algeria and Saudi Arabia

“…Machine Learning (ML) is defined as "the science of making computers act without being programmed." This process attempts to detect meaningful patterns between the inputs and build an autonomous model capable of describing these patterns without the need for human intervention (Bonello, Bredart, & Vella, 2018).…”