The government is testing a new system that will assist banks in assessing credit risks and the probability of frauds using big data analysis. The system is expected to help lenders, particularly rural and cooperative banks, tackle the issue of rising non performing assets (NPAs).
A credit rating model has been developed “that can assist the rural and cooperative banks to quantify risks under the big data context. Currently, rural and cooperative banks depend on judgement of the bank manager, resulting in high NPA,” according to an internal ministry note.
The project – sponsored by Ministry of Electronics and IT, has partners including the Reserve Bank of India, Bangalore-based IT firm Processware System and two cooperative banks.
A senior official, who did not wished to be named, said the project was undertaken with an objective that banks be helped to quantify risks associated with retail loans such as gold loan, personal loan and vehicle loans.
Under the project, a statistical and machine learning algorithmic models has also been developed to predict probability of default with an aim to improve NPAs. “A model has also been developed for predicting different types of frauds in the banking sector based on RBI guidelines,” the official added.
Additionally, a web enabled software is currently being tested which when implemented will assist the banks to easily adopt the models for credit rating, non-performing assets and fraud.”
As per the note, validation of the analytical models has been carried out using data from several banks.