There are a number of factors that can indicate the risk scoring and default prediction estimates of a given applicant.

Feeding multiple data sources into a custom model improves the accuracy of credit risk scoring and, more importantly, can be applied to fraud detection.