An Open Source Replication of a Winning Recidivism Prediction Model
We present results of our winning solution to the National Institute of Justice recidivism forecasting challenge. Our team, “MCHawks,” placed highly in both terms of accuracy (as measured via the Brier score), as well as the fairness criteria (weighted by differences in false positive rates between...
| Main Author: | |
|---|---|
| Contributors: | |
| Format: | Electronic Article |
| Language: | English |
| Published: |
2025
|
| In: |
International journal of offender therapy and comparative criminology
Year: 2025, Volume: 69, Issue: 5, Pages: 438-453 |
| Online Access: |
Volltext (lizenzpflichtig) |
| Journals Online & Print: | |
| Check availability: | HBZ Gateway |
| Keywords: |
MARC
| LEADER | 00000naa a22000002c 4500 | ||
|---|---|---|---|
| 001 | 1920046720 | ||
| 003 | DE-627 | ||
| 005 | 20250319054811.0 | ||
| 007 | cr uuu---uuuuu | ||
| 008 | 250319s2025 xx |||||o 00| ||eng c | ||
| 024 | 7 | |a 10.1177/0306624X221133004 |2 doi | |
| 035 | |a (DE-627)1920046720 | ||
| 035 | |a (DE-599)KXP1920046720 | ||
| 040 | |a DE-627 |b ger |c DE-627 |e rda | ||
| 041 | |a eng | ||
| 084 | |a 2,1 |2 ssgn | ||
| 100 | 1 | |a Circo, Giovanni |e VerfasserIn |0 (DE-588)1221227157 |0 (DE-627)1738481751 |4 aut | |
| 109 | |a Circo, Giovanni |a Circo, Giovanni M. | ||
| 245 | 1 | 0 | |a An Open Source Replication of a Winning Recidivism Prediction Model |
| 264 | 1 | |c 2025 | |
| 336 | |a Text |b txt |2 rdacontent | ||
| 337 | |a Computermedien |b c |2 rdamedia | ||
| 338 | |a Online-Ressource |b cr |2 rdacarrier | ||
| 520 | |a We present results of our winning solution to the National Institute of Justice recidivism forecasting challenge. Our team, “MCHawks,” placed highly in both terms of accuracy (as measured via the Brier score), as well as the fairness criteria (weighted by differences in false positive rates between White and Black parolees). We used a non-linear machine learning model, XGBoost, although we detail our search of different model specifications, as many different models’ predictive performance is very similar. Our solution to balancing false positive rates is trivial; we bias predictions to always be “low risk” so false positive rates for each racial group are zero. We discuss changes to the fairness metric to promote non-trivial solutions. By providing open-source replication materials, it is within the capabilities of others to build just as accurate models without extensive statistical expertise or computational resources. | ||
| 650 | 4 | |a open-science | |
| 650 | 4 | |a Prediction | |
| 650 | 4 | |a machine-learning | |
| 650 | 4 | |a Recidivism | |
| 700 | 1 | |a Wheeler, Andrew P. |e VerfasserIn |0 (DE-588)1170829775 |0 (DE-627)1040128602 |0 (DE-576)512607397 |4 aut | |
| 773 | 0 | 8 | |i Enthalten in |t International journal of offender therapy and comparative criminology |d London [u.a.] : Sage Publ., 1966 |g 69(2025), 5, Seite 438-453 |h Online-Ressource |w (DE-627)325163286 |w (DE-600)2034467-3 |w (DE-576)094533156 |x 1552-6933 |7 nnas |
| 773 | 1 | 8 | |g volume:69 |g year:2025 |g number:5 |g pages:438-453 |
| 856 | 4 | 0 | |u https://doi.org/10.1177/0306624X221133004 |x Resolving-System |z lizenzpflichtig |3 Volltext |
| 935 | |a mkri | ||
| 951 | |a AR | ||
| ELC | |a 1 | ||
| LOK | |0 000 xxxxxcx a22 zn 4500 | ||
| LOK | |0 001 4688069650 | ||
| LOK | |0 003 DE-627 | ||
| LOK | |0 004 1920046720 | ||
| LOK | |0 005 20250319043606 | ||
| LOK | |0 008 250319||||||||||||||||ger||||||| | ||
| LOK | |0 035 |a (DE-2619)KrimDok#2025-03-18#452A751DE58270F6FDDC46DD3DFF2B5F36A717DD | ||
| LOK | |0 040 |a DE-2619 |c DE-627 |d DE-2619 | ||
| LOK | |0 092 |o n | ||
| LOK | |0 852 |a DE-2619 | ||
| LOK | |0 852 1 |9 00 | ||
| LOK | |0 935 |a zota | ||
| ORI | |a SA-MARC-krimdoka001.raw | ||