Automation of Scoring Creditworthiness Using Optical Character Recognition and Probability of Default Calculation Algorithms

  • Aditya Khant ,Harvey Mudd College Claremont , California, USA (akhant@hmc.edu)
  • Paritosh Chandran ,NeuralTechSoft Goregaon, Mumbai, India (paritosh@neuraltechsoft.com)
  • Mahendra Mehta ,NeuralTechSoft Goregaon, Mumbai, India (mahendra@neuraltechsoft.com)

The automation of determining creditworthiness requires the machine to automatically gather and analyze financial data. Current methods require several man hours of data entry and analysis of data to assess a single case. This paper demonstrates the usage of state-of-the-art Optical Character Recognition (OCR) and Computer Vision (CV) techniques to gather data automatically from scanned documents in a format that can be used to process and analyze data further. This paper also details the novel automation of analysis of the extracted data to find the probability of default (PD) for a specific case. We find that this process reduces the turn around time of granting loans and minimizes human interactions and biases.

Please fill the following information to view the document.