Designing and explaining the model for predicting bankruptcy of companies using a deep learning model optimized with the whale meta-heuristic algorithm (WOA).

Articles in Press

Document Type : Original Article

Authors

1 میدان چمران - چمران ۳ - بن بست حسن مسعودی

2 Assistant Professor of Accounting, School of Management, Islamic Azad University, Khomein, Khomein Branch, Iran

3 Associate professor of Accounting Payame Noor University, Iran

10.22034/iaas.2023.392758.1452

Abstract

Today, businesses need to properly manage their resources and expenses in order to survive, in the competitive arena, the flexibility of companies has decreased drastically and this factor has caused them to not have the ability to react correctly and appropriately in different economic conditions and with the risk of bankruptcy. to face Predicting the bankruptcy of companies is one of the important topics that contribute to the success and continuity of companies. The purpose of this research is to design and explain the model of predicting the bankruptcy of companies using a deep learning model optimized with the whale meta-heuristic algorithm. It has been implemented on the data of 328 examples of listed companies including 246 healthy companies and 82 bankrupt companies in the period of 2012-2022. The financial ratios are the independent variables of this research, which were optimized using the whale meta-heuristic algorithm and extracted as one of the artificial intelligence models. The results showed that the ratios of operating profit to total assets, cash to net sales, Cash to total assets, cash to current liabilities, current liabilities to total assets have been the most effective variables in determining bankruptcy and in all evaluation criteria of classification models, fit function and area under the ROC curve of the whale algorithm compared to the swarm algorithm. The particles provided better performance.

Keywords

Main Subjects

Accounting and Auditing Studies

Articles in Press, Accepted Manuscript
Available Online from 19 February 2024

Files

Share

How to cite

Statistics