Machine Learning Analysis in Predicting Bankruptcy in Companies (Case Study of Manufacturing Companies Listed on the Stock Exchange)

This study aims to assess the effectiveness of machine learning in improving the accuracy of corporate bankruptcy prediction and eval_uate various machine learning models to determine the most effective model in predicting corporate bankruptcy. This study uses a case study with a descriptive analysis approach. The study population is manufacturing companies listed on the Indonesia Stock Exchange (IDX) in 2013-2023, using the Ohlson and Altman Z model datasets, each consisting of 1381 data. The hyperparameter tuning technique through GridSearchCV is used to determine the best parameters in forming the SVM model. The results show that the SVM model with the best parameters ('C': 10, 'kernel': 'linear') can classify Ohlson model data with 98.19% accuracy and Altman Z model data with 100% accuracy. The Random Forest model showed 94.11% accuracy in Ohlson model data and 97.31% in Altman Z model data. XGBoost model showed 97.70% accuracy in the Ohlson model data and 97.70% in the Altman Z model data. LSTM model showed 91.88% accuracy in Ohlson model data and 95.75% in Altman Z model data. The Altman Z model tends to be static and does not include external factors, while Ohlson's model is more complex but provides a more comprehensive picture of the company's health and prospects.