a lean and sustainable conscientious cannabusiness; target market? Black market.
-
Updated
Sep 2, 2020
a lean and sustainable conscientious cannabusiness; target market? Black market.
Predicting Recidivism and Analysis of National Corrections Data 1991 - 2014
Inspired by the National Institute of Justice's Recidivism Forecasting Challenge. It aims to refine recidivism forecasts using both person- and place-based variables to aid community supervision and reintegration efforts.
CMU 95791-Data Mining Group Project on Recidivism Forecasting
Predicting and characterizing recidivism in Colombia as part of a group project (Team 77) for DS4A certification.
Responsible criminal justice population risk analysis using NIJ recidivism data, SQLite, Python/R, subgroup diagnostics, threshold policy analysis, and Power BI-ready reporting.
XAI analysis on the COMPAS dataset for recidivism
A repository of various articles, abstracts, essays, and studies which analyze successful reentry strategies and report methods of avoiding recidivism.
ML-based feature selection + conditional discrimination mitigation with unified evaluation and outputs
NIJ Recidivism Forecasting Challenge — deep-learning model + 2026 revised findings: re-offending concentrates in a persistent high-prior-history cohort. Reproducible analysis included.
Data and analysis for 'Machine Bias'
Reoffending-risk prediction with a Neo4j knowledge graph and a Fairlearn audit on the COMPAS dataset. A small auditable architecture demonstration.
Fairness-aware recidivism risk assessment - calibrated predictions, SHAP explanations, conformal uncertainty, human-in-the-loop overrides, and a full audit trail.
Analysis of ~10 years of Connecticut DOC custody data: who is held pretrial and after sentencing, how long they stay, and how often they return. BigQuery SQL, pandas, Looker Studio.
Machine learning analysis of the COMPAS recidivism dataset examining predictive accuracy, algorithmic fairness, and behavioral risk factors associated with two-year reoffending.
R code to aggregate and merge features from different government sources, create flat file, perform relationship, prediction, and equity measures
Bureau of Justice Statistics — independent third-party profile of a public API surface, by API Evangelist. The Bureau of Justice Statistics (BJS) publishes information on crime, criminal offenders, victims of crime, and the operation of justice systems.
To associate your repository with the recidivism topic, visit your repo's landing page and select "manage topics."