Dissertations, Theses, and Capstone Projects
Date of Degree
6-2024
Document Type
Dissertation
Degree Name
Ph.D.
Program
Computer Science
Advisor
Feng Gu
Committee Members
Liang Zhao
Shuqun Zhang
Mengyan Dai
Subject Categories
Graphics and Human Computer Interfaces | Other Computer Sciences | Software Engineering
Keywords
agent-based modeling; crime prediction; crime analysis; simulation environment
Abstract
Crime risk evaluation and crime prediction using agent-based modeling (ABM) have gained popularity in the field of computational criminology in recent years. Traditionally, researchers rely on statistical methods and machine learning models to predict crimes using historical data. ABM generates macro-level crime patterns in a bottom-up fashion by simulating the daily behaviors of autonomous entities, such as citizens and offenders. ABM takes into consideration the non-linear interactions between agents under complex social contexts. Currently, the comprehensive usage of ABM for criminological theory testing and urban policy evaluations calls for a unified software framework. In this research, we introduce CARESim, an integrated simulation environment for crime analysis and risk estimation (CARE), to predict the risks of high-risk individuals in commit- ting street violent crimes. We focus on discussing three aspects of CARESim that benefit both criminological researchers and law-enforcement practitioners, including the construction, the validation, and the applications of CARESim. In model construction, there are several key components. The data conversion process transforms various kinds of data into functional model elements. The model environment provides the moving background based on the geographical information system (GIS) with different environmental layers stacked together. The conceptual model includes the behavioral design and categorization of different types of agents, the location selection system based on points of interest, and the risk evaluation process based on various criminogenic factors. The evaluation interface provides support for visual validation. Finally, the empirical experience of local experts helps formalize the conceptual model and overlook the whole validation process. In model validation, we present our three-phase face validation with domain experts, spatial pattern replication, and sensitivity analysis on key model parameters that lack empirical agreements. We demonstrate the use of CARESim by conducting experiments on the optimization of patrol strategy and variations of key factors, such as social networks, the neighborhood, the weather, and citizens’ behaviors and composition. Lastly, we evaluate the effectiveness of focused deterrence through randomized controlled trials using CARESim and real-life offenders’ data. In general, for the study of crime dynamics, the development of CARESim provides other researchers with a customizable simulation platform that can be efficiently adapted to other urban environments with guidance on data selection. Our experiments show the advantage of ABM in adjusting any theoretically devisable parameter for policy evaluations. The implementation of simulated focused deterrence offers guidance on individual-level crime prediction that boosts the confidence of stakeholders before applying it to actual field experiments.
Recommended Citation
Gong, Yifei, "Crime Prediction Using Agent-Based Modeling" (2024). CUNY Academic Works.
https://academicworks.cuny.edu/gc_etds/5738
Graduate Center users:
To read this work, log in to your GC ILL account and place a thesis request.
Non-GC Users:
See the GC’s lending policies to learn more.
Included in
Graphics and Human Computer Interfaces Commons, Other Computer Sciences Commons, Software Engineering Commons