Financial Distress Prediction Using an Artificial Neural Network Integrating a Two-Stage Feature Selection Method

Zijiao Zhang; Shiyou Qu; Chong Wu

doi:10.5755/j01.ee.35.4.30924

Authors

Zijiao Zhang School of Management, Harbin Institute of Technology, China
Shiyou Qu School of Management, Harbin Institute of Technology, China
Chong Wu School of Management, Harbin Institute of Technology, China

DOI:

https://doi.org/10.5755/j01.ee.35.4.30924

Keywords:

Financial distress prediction, Feature selection, Artificial neural network, Genetic algorithm, SHapley Additive exPlanations

Abstract

Financial distress prediction (FDP) is critical for companies, banks, and investors, and artificial neural networks (ANN) have been proven to be an efficient method for FDP. However, the “curse of dimensionality” in FDP not only increases the computational complexity, but also reduces the prediction accuracy. To solve this problem, this paper takes an ANN model as the basic classifier and presents a new two-stage feature selection method integrated with multiple filters and a wrapper method. The financial data of Chinese listed companies are applied for comparative analysis to verify the effectiveness of the constructed method. The results demonstrate that the proposed method achieves a smaller feature subset and better predictive effect than other methods, thus solving the “curse of dimensionality” more effectively and improving the accuracy. In addition, SHapley Additive exPlanations (SHAP) and Partial Dependence Plots (PDP) are employed to investigate the relative importance of selected features. Their results increase the credibility of the proposed model, giving users more confidence in using this “black box” model.

Author Biographies

Zijiao Zhang, School of Management, Harbin Institute of Technology, China

Zijiao Zhang is pursuing her PhD at the School of Management, Harbin Institute of Technology, China. Her research interests are in financial distress prediction, feature selection, and optimization algorithms. The recent research results are published in the journals Information Processing and Management, http://doi.org/10.1016/j.ipm.2022.102988, and Mathematical Biosciences and Engineering, http://doi.org/10.3934/mbe.2023924.

Shiyou Qu, School of Management, Harbin Institute of Technology, China

Shiyou Qu is a professor at the School of Management, Harbin Institute of Technology, China. He received his Ph.D. in Technical Economics and Management from Harbin Institute of Technology in 2003. He has published more than 20 peer-reviewed papers. His current research interests are in corporate finance and technological innovation theory.

Chong Wu, School of Management, Harbin Institute of Technology, China

Chong Wu is a professor at the School of Management, Harbin Institute of Technology, China. He received his Ph.D. in Fundamental Mathematics from Harbin Institute of Technology in 1998. He has published more than 70 peer-reviewed papers, including IEEE Transactions on Knowledge and Data Engineering, Journal of the Association for Information Science and Technology, Information Sciences, Knowledge-based Systems and Expert Systems with Applications. His current research interests are in prediction theory and applications.