I've developed a few models based on historical data to price folio loans. I built different models for loans with different status.
Loans that are current are priced based on historical default & prepayment behaviours and recent fico changes. Late loans are priced based on potential recoveries and are highly uncertain. My model couldn't price 'In Grace Period' loans because lack of historical data.
I've set up automatic process to price all loans once every week.
You can download my model output here
https://app.box.com/s/mtmeot4pxbdrznf3s6etw7gbjamvpro7feedbacks are welcome.
This could be interesting if there were additional columns with key parameters. I would suggest interest rate, term, payments made, and credit trend (or change in fico since origination, if you database that).
If you want to trade on it, you'll need to do it more than once a day. Mispriced notes are bought quite quickly