Student Name

Rizwan AnsariFollow

Degree

Master of Science in Data Science

Department

Department of Computer Science

Faculty/ School

School of Mathematics and Computer Science (SMCS)

Date of Submission

Spring 2023

Supervisor

Dr. Tahir Syed, Assistant Professor, Department of Computer Science, School of Mathematics and Computer Science (SMCS)

Co-Supervisor

Dr. Behroz Mirza, Head of Technical Program Management, Jubilee Life Insurance Company Ltd.

Keywords

Credit Risk, Model, Validation, Automation, Engine

Abstract

This project aims to explore the importance of risk scorecard models in the banking industry and the need for an automated credit risk model validation engine to ensure Discriminatory power, Significance, Multicollinearity and Stability of scorecards model which consist of nine different tests. Credit risk models are essential tools used by financial institutions to manage risk associated with lending and investment activities. To ensure the dimension mentioned earlier of these models, they must undergo a rigorous validation process. The Central Bank of UAE (CB UAE) Model Management Standards and Guidance and the UK Prudential Regulation Authority (PRA) Model Risk Management Principle provide guidelines on how to validate credit risk models effectively.

In this project, we propose a Credit Risk Models Validation Engine that aligns with the CB UAE Model Management Standards and Guidance and the UK PRA Model Risk Management Principle covering Quantitative validation. The engine automates the validation process of credit risk models, utilizing statistical techniques to ensure the validation results of the models with a GUI solution.

The proposed engine consists of three main modules: preprocessing, validation, and visualization. The preprocessing module cleans and transforms the input data to prepare it for EDA analysis. The validation module evaluates the model's performance against the expected results to assess model Discriminatory power, Significance, Multicollinearity and Stability, and the visualization module presents the validation results in an interactive and user-friendly manner.

The Credit Risk Models Validation Engine showcased exceptional performance, accuracy, reliability, and processing speed. Validation tests on two distinct datasets affirmed the engine's effectiveness and robustness in evaluating credit risk models health and results. With its ability to accurately validate models and offer valuable insights, the engine serves as a powerful tool for credit risk management decision-making.

Document Type

Restricted Access

Submission Type

Research Project

Loading...

Media is loading
 

The full text of this document is only accessible to authorized users.

Share

COinS