Machine learning methods in capital structure decisions; application in bist

This study was conducted to create a model based on machine learning methods to determine and develop firm-specific factors that significantly affect the funding decisions of firms. The study consists of firms listed in the forest, paper and printing index of Borsa Istanbul between 2014-2023, and their capital structure decisions were analyzed with machine learning methods. As a result of the analyzes made using machine learning methods, it was determined that the performance values of the model created with the XGBoost method and the Random Forest model, in particular, performed better than the AdaBoost and SVM models. It was seen that the models created with the XGBoost and Random Forest methods were a better option for determining corporate capital structures. It was understood that the SVM model had the worst performance value.