Bankruptcy Forecasting of Indian Manufacturing Companies Post the Insolvency and Bankruptcy Code 2016 Using Machine Learning Techniques

Purpose: Bankruptcies have increased dramatically in recent years. The manufacturing industry is one of the most important contributors to the country’s Gross Domestic Product (GDP). The GDP of a country reflects its development and progress. More and more bankruptcies in the manufacturing business will have a significant influence on the country's GDP. The primary goal of this study is to conduct a comparative analysis of numerous bankruptcy predictive models in order to recommend the optimal model with the highest accuracy for bankruptcy prediction. Methodology/Approach: This research employs a number of machine-learning forecasting approaches. Logistic Regression, Decision Tree, Artificial Neural Networks (ANN), and Random Forest are the machine learning techniques employed in this paper. A comparison study is conducted with and without Principal Component Analysis (PCA). A total of 15 financial factors were identified from prior studies, and a comparative analysis was conducted with those variables. From 1 April 2017 to 31 March 2020, the Insolvency and Bankruptcy Board of India (IBBI) database is used to collect information on bankrupt companies. Data for the previous three years is gathered from the annual reports of 70 enterprises (35 bankrupt, 35 non-bankrupt). Contribution: This paper adds to the existing research on bankruptcy. There is relatively limited research on bankruptcy prediction after the implementation of the Insolvency and Bankruptcy Code (IBC), 2016. Most studies on bankruptcy prediction in India used logistic regression or ANN because of their widespread use and good accuracy. In India, very few research used decision tree-based methodologies to forecast bankruptcy. This research, on the other hand, contributes to decision tree-based studies and they are showing more accurate results as compared to ANN or logistic regression. Limitations: One of the major limitations of this paper is that it mainly considers financial variables for research. Recent research has considered not just financial variables, but also corporate governance indicators and macroeconomic variables. Another disadvantage is that this report primarily focuses on the manufacturing industry, thus bankruptcy research in other industries is required.