APPLICATION OF ARTIFICIAL NEURAL NETWORK MODEL FOR CREDIT SCORING IN PEER-TO-PEER LENDING

• Main Author: Tejasatya Lubis, Deryan
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/50349
• Institution: Institut Teknologi Bandung

Credit or loan is a financial product that allows someone to borrow money from other parties and pay it back within the specified time period. Currently there are many ways for someone to obtain a loan, one of them is through peer-to-peer lending companies. Peer-to-peer lending companies bring together borrowers and lenders directly through a platform. The presence of peer-to-peer lending companies can provide a more inclusive access to loans for those who do not have a bank account(the unbanked). However, peer-to-peer lending companies are dealing with higher risks compared to conventional loans like those of banks. This happens because it’s usually harder for the companies to assess a borrower’s risk in peer-to-peer lending scheme. On the other hand, data analysis technologies such as machine learning models are now available. Using machine learning, we are able to find patterns that are difficult to understand by humans. Therefore, in this study, the author implements artificial neural network model for credit credit scoring in peer-to-peer lending. Modeling will be carried out using CRISP-DM, using a public dataset from Lending Club, and Python 3 programming language for programming and machine learning modelling. In this study, an artificial neural network model have been obtained with an 95.75% accuracy, 90.89% precision, 84.97% sensitivity, and 87.83% F1 Score.