Credit Risk Analysis in Peer-to-Peer Lending System

摘要

This research paper aims to analyze the credit risk involved in peer-to-peer (P2P) lending system of “LendingClub” Company. The P2P system allows investors to get significantly higher return on investment as compared to bank deposit, but it comes with a risk of the loan and interest not being repaid. Ensemble machine learning algorithms and preprocessing techniques are used to explore, analyze and determine the factors which play crucial role in predicting the credit risk involved in “LendingClub” publicly available 2013-2015 loan applications dataset. A loan is considered “good” if it's repaid with interest and on time. The algorithms are optimized to favor the potential good loans whilst identifying defaults or risky credits.