May
25
2011

About “blended” scoring models

In my last post we discussed how the best predictive performance comes from using the consumer data on the business owner and blending it with the commercial data on the business.  But how does the scoring model work using both sets of data?  Unfortunately it is not as simple as half of the score is driven by the consumer data and half by the business data.  Advanced statistically processes are used to identify the relationship between data elements and the best prediction of future performance.  Thus we have often seen very high consumer scores on the business owner and yet the blended business score is low because of significant derogatory commercial data.  The use of both sets of data always means a better identification of risk.