Prediksi kebangkrutan perusahaan menggunakan metode klasifikasi: Studi kasus pada industri

Abstrak

Corporate bankruptcy prediction is a crucial aspect of the financial sector because it can significantly affect investors, creditors, company management, and other stakeholders in making strategic decisions. This study aims to develop an accurate bankruptcy prediction model using the XGBoost algorithm optimized through hyperparameter tuning. In addition, the Synthetic Minority Over-sampling Technique (SMOTE) was applied to address data imbalance by increasing the representation of the minority class. The model was also tested on two large-scale datasets (Taiwan and US) to assess its performance consistency and generalization capability. The results show that XGBoost with hyperparameter tuning achieves the best performance, with an accuracy of 98.94%, a precision of 0.98, and a recall of 1.0. Furthermore, the model demonstrated stable performance without indications of overfitting. These findings confirm that XGBoost with hyperparameter tuning can provide accurate, consistent, and reliable bankruptcy predictions and have strong potential for broader industrial implementation and larger-scale applications.