Novel Framework for Forecasting Financial Distress Using Advanced Machine Learning Technique

Abstract

Financial hardship is a serious problem that impacts people, businesses, and economies. It frequently produces severe outcomes, including job loss, bankruptcy, and economic downturns. Financial distress refers to an organization's inability to meet its financial obligations, which can stem from inadequate financial management, recessions, unforeseen costs, or significant market shifts. The study aims to establish an advanced machine learning (ML) technique for forecasting financial distress in companies. The study proposed a novel resilient pinecone algorithm-mutated extreme boosting (RPA-XGBoost) model to predict a company’s financial distress. The RPA-XGBoost model aims to determine whether a business is in a state of financial distress. RPA-based methods for parametric improvement in the financial and XGBoost models receive the financial data as input and predict financial distress. The study utilizes the financial metrics of various companies along with indicators of financial distress. The data was preprocessed using data cleaning and normalization for the obtained data. The results demonstrate traditional algorithms outperform the proposed method in terms of the higher instances precision of 0.99, accuracy of 98.71, f1-score of 0.989, recall of 0.992 and time complexity of 20.12. In conclusion, this study presents a novel framework for forecasting financial distress that leverages advanced ML techniques to enhance predictive accuracy and reliability.