Date of Graduation

12-2020

Document Type

Thesis

Degree Name

Master of Science in Statistics and Analytics (MS)

Degree Level

Graduate

Department

Graduate School

Advisor

Jyotishka Datta

Committee Member

Grant Drawve

Second Committee Member

Samantha Robinson

Third Committee Member

Giovanni Petris

Keywords

resource allocation, meta model, statistical models, methodoogy

Abstract

Proper allocation of law enforcement agencies falls under the umbrella of risk terrainmodeling (Caplan et al., 2011, 2015; Drawve, 2016) that primarily focuses on crime prediction and prevention by spatially aggregating response and predictor variables of interest. Although mental health incidents demand resource allocation from law enforcement agencies and the city, relatively less emphasis has been placed on building spatial models for mental health incidents events. Analyzing spatial mental health events in Little Rock, AR over 2015 to 2018, we found evidence of spatial heterogeneity via Moran’s I statistic. A spatial modeling framework is then built using generalized linear models, spatial regression models and a tree based method, in particular, Poisson regression, spatial Durbin error model, Manski model and Random Forest. The insights obtained from these different models are presented here along with their relative predictive performances. These inferential tools have the potential to aid both law enforcement agencies and the city in properly allocating resources required for such events.

Share

COinS