Description of original award (Fiscal Year 2018, $500,000)
The applicant proposes to develop an expert system to guide homicide investigation praxis by using machine learning (ML) techniques. The expert system algorithm will be trained using 4,000 police case files (Murder Books) for murders committed between 1990-2010 in Los Angeles, CA. Murders involving gangs will receive particularly attention.
These murder books contain a plethora of information/data sets (e.g., police logs, detective notes, forensic results, and witness/suspect information). In addition to development of the expert system algorithm, the applicant will use the information/data sets contained in the murder books to determine: (1) how homicides may differ from one Police Bureau to another; (2) examine how gang-related homicides and shootings differ from other types of homicides and shootings in terms of clearance rates; and (3) measure the degree to which predictors of gun homicide and shooting incidents are similar and different. This effort will culminate with a one-year evaluation of the active deployment of this tool in the LAPDs South Bureau.
To evaluate the impact of adopting this tool on homicide clearance, the applicant will employ an Interrupted Time Series (ITS) research design to examine trends in homicide clearance, homicide conviction, and homicide incidents monthly from 2000 through November 2020. (ITS is a type of quasi-experimental design that is robust to many of the threats to internal validity.) These data will be analyzed using Segmented Regression, which allows for a clear interpretation of the impact of the intervention. While there will only be a limited number of post-intervention periods to examine, this initial analysis can inform future research.
Note: This project contains a research and/or development component, as defined in applicable law," and complies with Part 200 Uniform Requirements - 2 CFR 200.210(a)(14).