Assessing the impact of interventions on retaliatory violent crimes using Hawkes models with covariates
- Author(s): Wang, Huanchen
- Advisor(s): Schoenberg, Frederic P
- et al.
Recent studies suggest the effective application of Hawkes point process models to describe the self-exciting nature of gang crimes. We focus on applying Hawkes point process models with spatial covariates incorporated in the background rate to assess the City of Los Angeles Mayors Office Gang Reduction and Youth Devel- opment (GRYD) intervention program. The data includes all officially recorded crimes that took place in South Los Angeles during 2014-2016. The fitted model is compared to a previous study (Brantingham et al. 2018) conducted on the same topic using Hawkes point process model with smoothing kernel background rate. The GRYD intervention program is found to be involved at locations with highly clustered crimes, which has led to difficulty in assessment. After accounting for this differential placement of GRYD events, the fitted model shows that the GRYD intervention program results in effective reduction of gang crime retalia- tions, although the results are not statistically sigificant.