Skip to content

ROCeey/Credit-Default-Prediction

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

10 Commits
 
 
 
 
 
 

Repository files navigation

Application of AI Models for Credit Risk Assessment in P2P Lending

image

Project Rationale

Because bank lending is the most common source of external funding for many Small and Medium Enterprises and entrepreneurs, there is a growing demand for loans to people and SMEs. People and businesses require loans for a variety of reasons, including cash emergencies, debt consolidation, and the start-up and/or expansion of a company. A substantial number of individuals and organizations in Nigeria are unable to obtain credit, and SMEs face similar challenges due to issues such as a lack of collateral, poor record-keeping, poor creditworthiness, and weak project proposals. For this reason, peer-to-peer lending platforms, which allow people and businesses to borrow and lend money without the involvement of a traditional financial institution have evolved over time.

However, P2P lending platforms are linked to high credit risks since borrowers have a high tendency to default, either by not repaying disbursed loans on time or at all. For P2PL organizations, managing loan default risk is crucial and it is advisable for them to deploy models that can accurately predict loan default tendencies to help mitigate these risks.

Building algorithms that identify specific consumer categories and allocate weights based on risk levels, in my opinion, would assist an investor in lessening the risks associated with P2P lending. This research will analyze how Artificial Intelligence (AI) models can help Financial Technology Institutions in Nigeria improve the P2P lending process by reducing the chances of borrowers not meeting their repayment commitments when they are due.

Project Aim

The overall aim of this project is to create a predictive loan risk model that will guide P2P lending platform investors to make informed decisions about borrowers with high default risks based on certain characteristics.

Research questions

The major questions stated below should represent the bulk of this research's primary findings to meet the project's main goals and objectives. Among them are:

  1. What are the key contributing factors that increase a borrower’s likelihood of defaulting on P2P lending platforms?
  2. Is the risk of loan default larger for first-time loans or repeat loans?
  3. How accurately is the model performing to predict the probability of borrowers that will default?

Approach

We will give consideration to the impact of first time and repeat loans in predicting credit risk default using supervised machine learning. We will also consider a deep-learning approach to the problem as well.

Models functionality

Our models should be able to make calculated conclusions based on the likelihood of repaying the loan at the time of application, determining whether or not the client should be granted a loan.

About

Predict loan default in a peer-to-peer lending settings

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published