Forests identify built-environment associations in recorded drug-crime patterns
Source check: Source checked. Not independent scientific confirmation.
Finding: Random forests analyzed 52,567 reported drug-related incidents in Chicago during 2016–2019. Vacant lots and buildings ranked highly in models distinguishing recorded heroin and synthetic-drug clusters. These are neighborhood-level associations in police records, not causal effects or measures of individuals’ behavior.
Technology: Space-time random forest; guided regularized random forest; partial dependence
Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- Forest ensembles learn nonlinear feature combinations; corrected importance and partial dependence characterize the source-specific observed spatial associations.
- Test and result
- Forest feature selection and importance analysis compared recorded cluster patterns with built-environment, demographic and lagged variables.
- Limitation / still unknown
- Policing practices were unavailable and may influence recorded patterns; ecological associations do not identify individual risk or prove neighborhood causes. Temporal-window selection used the studied series. No treatment-service or policing benefit was tested.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC8021089/
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 21 Nov 2026. Review scope: Indexed primary data/model/feature-selection Methods, Results and limitations; exact primary front date/affiliations and fresh primary notice metadata..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
Editorial assessment: Impact 1/5 · Validation 2/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.