Cognerium Conducts Case Study Highlighting the Credit Risk Model’s Value to Banks

Cognerium Logo

Cognerium’s Rohit Garg outlines how the Credit Risk Model provides risk management solutions to banks using AI/ML

BOCA RATON, FLORIDA, UNITED STATES, January 19, 2021 / — A recent case study has been conducted by Rohit Garg, Chief Data Scientist at Cognerium to illuminate how the Credit Risk Model has been built and what value it can provide to clients. In this study, the client, a credit risk team of a mid-sized bank, wanted to predict the riskiness of applicants obtaining secured and unsecured loans for their personal lending business.

Banks are tasked with determining whether or not they should move forward with an individual’s application for a loan. This decision is ultimately based on the client’s evaluation of whether the individual can reliably repay the loan and principal interest.

Lending Ally is a platform developed by Cognerium consisting of multiple credit scoring models developed using multiple AI/ML models. These include the Logistic regression model, Decision trees, Random forest, Gradient boosting, and Artificial neural network. The data scientists at Cognerium provided this solution to the client to conduct their evaluation of the applicants.

In order to enact the study, the client provided the data composed of credit and personal information of a group of customers. The data set includes details on default payments, demographics factors, credit data, and history of payment.

Various statistics were then compared to evaluate the Lending Ally models. First, the scientists looked at accuracy, which measures the percentage of the correctly classified population to the total population. Next, they looked at the AUROC. The AUC or ROC curve shows the proportion of true positives (defaulter is correctly classified as defaulter) versus the proportion of false positives (non-defaulter is wrongly classified as a defaulter). Finally, the KS, which measures the maximum separation between the good population and the bad population, was compared. Based on accuracy, the algorithms are arranged in decreasing order.

Overall, Cognerium’s case study determined that the solution helped the client (Bank) conduct better risk management. The Lending Ally successfully identified the best AI/ML model for the client. With this substantial improvement in accuracy, the bank can increase the true positive rate by correctly identifying the applicant that will default; therefore, reducing the losses. Simultaneously, the bank can increase the true negative rate by correctly identifying the applicant that will not default, thus increasing the profits.

When asked to report the value the bank realized using the Credit Risk Model for loan applicants, Rohit responded, “When we ran the Credit Risk Model, on loan applicants data and macro-economic data from the past year, the model returned 41% of the loan portfolio with high-risk. This model can help banks understand each loan applicant’s risk-level and help banks to correctly position the borrowing terms and conditions to avoid losses. The same model can run on active loans to determine future risk of borrowers, taking any variation of market conditions into account.”

To learn more about Cognerium and the Credit Risk Model, please visit

About Cognerium

Cognerium is a Florida-based financial technology company specializing in AI-enabled Digital Transformations for banks and credit unions. With a vision to build intelligence that can assist humans in solving complex problems, Cognerium built the world’s first AI operating system for banking, financial services, and insurance (BFSI). The company’s services encompass business growth, credit risk and compliance, data strategy governance, and digital experience.

Rohit Garg
Visit us on social media:

Source: EIN Presswire