EdgeScore

EdgeScore is a proprietary machine learning-based risk score trained on consumer permissioned bank transaction data to be predictive of early term defaults (such as first payment default).
The score was developed using data from a variety of personal lending products and trained on targets including First Payment Default.

EdgeScore is output on a range of 500 to 750, with decreasing likelihood of default as the score increases.
Accounts that do not have sufficient data for reliable scoring will receive a 200 score.

Adverse Factors

Every scored record will also contain up to five (5) adverse factors.

Adverse Factor	Related Feature Categories
Excessive obligations in relation to income	Loan Activity, Income Activity, Balance Trends
Income insufficient for the amount of credit requested	Income Activity, Balance Trends
Limited credit experience	Account Activity, Account History
Temporary or irregular employment	Income Activity
Unable to verify income	Income Activity
Unable to verify employment	Income Activity
Unable to verify credit references	Loan Activity, Debt Activity

Odds Table

Band	Good vs Bad Odds	Cumulative Good	Cumulative Bad	KS
500 - 550	0.333	0%	0%	0%
550 - 560	0.869	0%	0%	0%
560 - 570	0.903	0%	2%	1%
570 - 580	1.094	1%	5%	4%
580 - 590	1.487	3%	11%	8%
590 - 600	1.861	5%	18%	13%
600 - 610	2.455	10%	28%	18%
610 - 620	2.817	16%	39%	23%
620 - 630	3.653	25%	51%	26%
630 - 640	4.593	36%	63%	27%
640 - 650	5.84	50%	75%	25%
650 - 660	7.082	64%	85%	21%
660 - 670	10.423	79%	92%	14%
670 - 680	12.127	89%	97%	8%
680 - 690	14.805	95%	99%	4%
690 - 700	17.164	98%	100%	2%
700 - 710	24.949	99%	100%	0%
710 - 720	26.538	100%	100%	0%
720 - 750	27.121	100%	100%	0%
	5.063	100%	100%