Data Oriented Financial Analysis using Machine Learning Methods

Financial analysts perform balance sheet adjustment before applicants' credibility scores are calculated in the assessment of commercial and corporate loan applications. The analysts usually go through financial documents manually and do reductions, additions or movements of balances in accounts. It causes waste of time and labor for financial institutions. This paper presents a solution model that detects balance sheet items to be adjusted in order to reduce costs, accelerate the balance sheet adjustment process by helping financial analysts and minimize the human error. This paper contributes to the literature by proposing a new feature set that can detect balance sheet items to be adjusted. The proposed solution model and feature set were tested and the results show that Stacked Generalization model, Random Forest as meta-learner and LightGBM, XGBoost and CatBoost as base learners, is the top performer model with the new feature set. The dataset used in experiments is obtained from Yapı Kredi, one of the largest banks of Turkey.