Year of Graduation
Predicting the Loan Default Probability Using Credit Bureau Data via Modern Machine Learning Methods
This work compares the quality of different models on the task of predicting default probability. The key novelty of the work is the use of a convolutional neural network for predictions, which is an algorithm typically employed in text or image analysis. This research also includes an overview of the most popular default probability prediction methods as well as a description of the basic principles of scoring. The data for training is taken from a kaggle.com competition. The results of the research can be used by companies to enhance their scoring algorithms.