Alternate Credit Score
Validation of Intent, Ability & Creditworthiness of Unbanked & Under-banked Customers using Machine Learning, NLP and Spark ecosystem
Our client is a leading financial technology company dedicated to creating financial technology for underserved markets. They are committed to providing the latest financial technology systems to help companies grow, become profitable, and remain profitable. They are currently developing modern, state of the art technology for lending, risk assessment, remittance, and investment spaces.
The main aim of this project is to develop a Risk Assessment Software System which can predict the risk assessment profile / creditworthiness of an organization / individual, determine future risk predictions and provide actionable information that businesses can use. The system must be able to access the financial information or keywords from various available sources of credit information including Social media.
Traditional credit scoring agencies has been providing great outcomes for the financial service sector for many years. Some of the underlying factors to help them determine credit scores included the below:
- Payment History
- Amount owed
- Credit history
- Duration & frequency of availing new credits
For the unbanked / under-banked / underserved customers, the above-mentioned parameters may not be available to arrive at a credit score using traditional algorithms.
- Estimate the credit score for the customers who is not having any financial identity. The system helps to take decision on recommending loan or not.
- To collect the data points from various sources available sources
- To analyse the data points from various sources and construct a structured dataset for further processing.
- Pick the appropriate data points which has the major influence on the credit score.
- To derive the credit score which access the risk level in lending the loans.
- The new credit scoring system has the potential to expand logical and responsible access to loan products.
- These can provide ample data points to the lenders that the borrower has the intent, ability and the creditworthiness to repay the loan successfully
- The lending institutions can also strengthen their decision making and alleviate the risks associated with people who might be rated creditworthy by the traditional methods and yet contributes to the NPA.
The derived credit score is more accurate than the traditional methods as it uses statistical model to assign a risk score to a credit application and it is a form of Algorithm, based on predictive modelling that assesses the likelihood of a customer defaulting on a credit obligation.
- Machine Learning
- Python 3.6
- Django Framework
- Enhanced Creditworthiness.
- Credit access made possible for the under-banked.
- Uses real-time data.
- Existing borrowers can get a better deal.
- Boosts the underwriting process.