n order to cater to the needs of the unbanked population, a simple approach would be to continue the existing infrastructure but giving relaxation to people with low-incomes. This would be helpful for unbanked population to have access to loans. However, this solution isn’t ideally feasible for the banks on the practical grounds.
Keeping all the aforementioned points in mind, our AI team formulated a criterion resembling to the one utilized by the banks but modified to make it suitable for the unbanked population.
Our team devised a machine learning model that predicts the credibility of a unbanked borrower while assessing his mobile phone recharge, grocery purchases, utility bills, online behavior and social data from Facebook, Instagram, Twitter, etc. At the moment, the model is providing a remarkable accuracy and due to latest feature engineering techniques, we’re able to operate very fine on small data.
