I have spent some time reviewing and revamping my loan to listing match algorithm.
Using yesterday’s data I am am able to match 7575 out of 7591 loans ~ 99.8%.
The 16 remaining loans look very similar and have creations dates within a day or two of each other.
Here is my algorithm (each pass finds only unique matches, once a match is found that listing can no longer be used in subsequent passes):
Pass #1: (7339 rows matched)
Match amount, lenderrate, creditgrade, DTI
loan.originationdate>listing.enddate
Pass #2: (224 rows matched)
Match amount, lenderrate, creditgrade, DTI, Group
loan.originationdate>listing.enddate
Pass #3: (12 rows matched)
Match amount, lenderrate, creditgrade, DTI
loan.originationdate>listing.enddate
(loan.originationdate-45)<=listing.enddate
If you would like my sql for personal use to add a listingkey to the loan table it is for sale for a $5 PayPal donation.
I am still not sure of the legal or moral arguments around publishing the sql script to the world. So you must use it for personal use only. Don’t share. Don’t publish. Your internal use only for analyzing loan opportunities. Don’t attribute specific loan/listing correlation to RateLadder for any purpose without prior written permission. No Warranty. No promise to be accurate. Past success is no guarantee it will work in the future.
If you liked this article, vote for it on del.icio.us and stumbleupon.Categories:
Prosper.com, Tools
Related Articles Related Stores
0 comments ↓
There are no comments yet...Kick things off by filling out the form below.
Leave a Comment