Named after the Norse God of gold and foreknowledge, our Credit Risk models will help you increase loan pass rates and reduce defaults by up to 15%
Now (Q2 2019)
The prediction model is accessed through a RESTful JSON API. Additional data is uploaded as a CSV file.
Our Credit Risk models look at predictive features in a Loan Application. By utilising near real-time, population relevant data sets from credit bureaus such as Experian, our GLM models can give significantly higher recall rates compared to a normal score card.
A credit scorecard is a lookup table that maps specific characteristics of a borrower into points. The total number of points becomes the credit score.
Like other credit scoring models, credit scorecards quantify the risk that a borrower will not repay a loan in the form of a score and a probability of default.
The problem with using out of date historic data to build either a risk model or a score card is that borrowers can suddenly go from having a perfect credit score to then going into default. By using our algorithms and looking for outliers and using what we call "smart data" such as the length of the application for a loan.
By running our algorithm on Lending Club data (where the outcome is known) we have shown our algorithms are more accurate than a standard score card.