Under the Award Number, this submission to the National Institute of Justice’s (NIJ’s) Recidivism Forecasting Challenge is entitled the “Winning Paper” of the Challenge.
The goal of the Recidivism Challenge is to improve the ability to forecast recidivism using person- and place-based variables to improve outcomes for those serving a community supervision sentence. NIJ hopes through the Challenge to encourage discussion on the topics of reentry, bias/fairness, measurement, and algorithm advancement. In addition to the Challenge data provided, NIJ encourages contestants to consider a wide range of potential supplemental data sources that are available to community corrections agencies to improve risk determinations, including the incorporation of dynamic place-based factors along with the common static and dynamic risk factors. NIJ is interested in models that accurately improve risk determination. In this winning project, the author analyzed and predicted the likelihood of recidivism using profile data. This was accomplished using Logistic regression, Random Forest Classifier, XGBoost. LightGBM, and Catboost algorithms accompanied by evaluated performance. The project was divided into six main parts, which involved exploratory data analysis, feature engineering, model building, model evaluation, feature importance, and inference. The Challenge uses data from the State of Georgia about persons released from prison to parole supervision for the period January 1, 2013 through December 31, 2015. The data include individual- and place-based variables that capture the supervision case information, prison case information, prior Georgia criminal history, prior Georgia community supervision history, Georgia board of pardons and parole conditions of supervision, and supervision activities.
Downloads
Similar Publications
- Raman spectroscopic signature of vaginal fluid and its potential application in forensic body fluid identification
- Insights into turning points from the perspective of young people with out-of-home care experience: events, impact and facilitators of change
- Two-Stage Approach for the Inference of the Source of High-Dimension and Complex Chemical Data in Forensic Science