Modelování kreditního rizika peer-to-peer půjček

Abstract

The aim of this thesis is to build two scoring models for peer-to-peer loan applicants using different methods and then compare their classification performance. The selected methods are logistic regression and classification tree. The first part is devoted to the principles, models, history, regulation and risks of peer-to-peer lending. The second part describes the methods used. In the third part, two scoring models are developed and then their classification performance is compared. Based on comparisons using the overall accuracy and area under the curve indicators, the logistic regression performs better than the classification tree.