Case Study: BFSI
Our Credit Risk Modelling can prevent potential loss of 1.2 Million Dollars within the first year
Existing solutions of in-house teams in credit risk modelling or predictions are not enough and the result could be a high default rate within both retail and corporate defaulters.
Abzooba identified increase in retail defaulters by 8.3% on an average of last 5 years and a whopping 13.25% for corporate defaulters and it goes to indicate that existing methodologies are not equipped enough to minimize the number of defaulters.
Implementation of Apache Spark as a crawling infrastructure to collect continuous data from LinkedIn provided us details of Names, current organization and contact supplemented by banking database to arrive at a staggering 92% accuracy within an year.
Our NLP tool xpresso ascertained accurate customer credibility factors from a varied recommendation text connections. The output of xpresso along with other KPIs such as job rotation, years of experience, frequency of job change and key skill endorsements to name a few was instrumental in adding explanatory variables in the Final Credit Scoring Model and the defaulters were correctly identified through a machine learning based logical regression.
Decrease in the rate of defaulters by 3.23%, thereby preventing a potential loss of $1.2 Million in 1st year itself.