Abstract:There has been intensive research from academics and practitioners regarding models for predicting bankruptcy and default events, for credit risk management. Seminal academic research has evaluated bankruptcy using traditional statistics techniques (e.g. discriminant analysis and logistic regression) and early artificial intelligence models (e.g. artificial neural networks). In this study, we test machine learning models (support vector machines, bagging, boosting, and random forest) to predict bankruptcy one year prior to the event, and compare their performance with results from discriminant analysis, logistic regression, and neural networks. We use data from 1985 to 2013 on North American firms, integrating information from the Salomon Center database and Compustat, analysing more than 10,000 firm-year observations. The key insight of the study is a substantial improvement in prediction accuracy using machine learning techniques especially when, in addition to the original Altman’s Z-score variables, we include six complementary financial indicators. Based on Carton and Hofer (2006), we use new variables, such as the operating margin, change in return-on-equity, change in price-to-book, and growth measures related to assets, sales, and number of employees, as predictive variables. Machine learning models show, on average, approximately 10% more accuracy in relation to traditional models. Comparing the best models, with all predictive variables, the machine learning technique related to random forest led to 87% accuracy, whereas logistic regression and linear discriminant analysis led to 69% and 50% accuracy, respectively, in the testing sample. We find that bagging, boosting, and random forest models outperform the others techniques, and that all prediction accuracy in the testing sample improves when the additional variables are included. Our research adds to the discussion of the continuing debate about superiority of computational methods over statistical techniques such as in Tsai, Hsu, and Yen (2014) and Yeh, Chi, and Lin (2014). In particular, for machine learning mechanisms, we do not find SVM to lead to higher accuracy rates than other models. This result contradicts outcomes from Danenas and Garsva (2015) and Cleofas-Sanchez, Garcia, Marques, and Senchez (2016), but corroborates, for instance, Wang, Ma, and Yang (2014), Liang, Lu, Tsai, and Shih (2016), and Cano et al. (2017). Our study supports the applicability of the expert systems by practitioners as in Heo and Yang (2014), Kim, Kang, and Kim (2015) and Xiao, Xiao, and Wang (2016).

The Application Of Kalman Filtering To Corporate Bankruptcy Prediction

A Financial Distress State Space Dynamic Model Based on Kalman Filtering

Applications of Kalman Filtering in Time Series Prediction

Towards augmented kernel extreme learning models for bankruptcy prediction: Algorithmic behavior and comprehensive analysis

Application of Kalman Filter in Stochastic Differential Equations

Business Failure Prediction Using Exponential Smoothing Forecasting and Pattern Recognition

Predicting corporate bankruptcy using the framework of Leland-Toft: evidence from U.S.

Application of kalman filter in time series software reliability growth model

Online chaotic time-series prediction with the derivative-free extended Kalman filter

Survey, classification and critical analysis of the literature on corporate bankruptcy and financial distress prediction

Machine Learning in Financial Crisis Prediction: A Survey

Bankruptcy Prediction Using Kriging Modeling

Bankruptcy Prediction Using Multiple Intelligent Agent System Via A Localized Generalization Error Approach

Financial Data Analysis Using Expert Bayesian Framework For Bankruptcy Prediction

Machine learning models and bankruptcy prediction

A hybrid financial analysis model for business failure prediction

Using partial least squares and support vector machines for bankruptcy prediction

Generalized Extended Kalman Filter For Prediction Of Chaotic Time-Series With Intermittent Failures

Bankruptcy Prediction Using SVM and Hybrid SVM Survey

Machine Learning Methods of Bankruptcy Prediction Using Accounting Ratios