Abstract

Globally increasing criminal activity, such as E Crimes, presents a serious threat to both economic growth and social welfare. Nevertheless, Dubai is a safety haven in the middle of this trend, as seen by the significant decline in major and non-alarming crimes during the first quarter of 2023. This study uses an examination of the large crime dataset (about 34,567 events) from the Dubai Police Department that spans 2019 to 2021. Several main goals are to be achieved by this work by using geospatial methods, predictive modelling, correlation investigations, and exploratory analysis. First of all, it looks for complex trends in the socioeconomic, geographical, and temporal aspects of Dubai's criminal scene. Later, working with legislators and law enforcement organizations, strategic interventions will be developed to reduce e-crime rates via use of predictive intelligence. The increase in e-crime, driven by the widespread use of smartphones and the internet, which has given rise to cyber risks like hacking, identity theft, and online fraud, is of special concern to the Dubai Police. Understanding that contemporary crime is changing, initiatives are being made to raise cybersecurity knowledge and monitor regional threat trends. The main objectives of this work are to develop machine learning classifier models to accurately predict e-crime behaviour, analyse demographic, temporal, and economic trends in e-crime statistics, and offer practical suggestions for resource allocation and crime reduction techniques. It is projected that e-crime rates in Dubai will drop by 20% by 2025 by leveraging the potential of data-driven regulations, therefore creating a safer and more secure environment for its residents. The study approach uses the CRISP-DM analytics paradigm and includes stages including business understanding, data understanding, data preparation, modelling, assessment, and implementation. Even with their inherent drawbacks—depending on official datasets, unpredictable social and economic forces, and the haziness of projections—modern analytics have enormous potential to support Dubai's safety efforts. In the study, the Logistic Regression model, utilizing 46 predictor fields, achieved an impressive accuracy of 85.871%, with an AUC of 0.932 and precision, recall, and F1-measure scores of 0.855. More detailed statistics must be included, model alarms must be integrated with surveillance systems, and models must be updated often to identify new patterns. A safer Dubai is possible by a coordinated effort to get beyond these limitations and significantly improve the effectiveness and efficiency of crime prevention methods.

Library of Congress Subject Headings

Crime--United Arab Emirates--Dubai; Cyberterrorism--United Arab Emirates--Dubai; Machine learning; Predictive analytics

Publication Date

5-2024

Document Type

Thesis

Student Type

Graduate

Degree Name

Professional Studies (MS)

Department, Program, or Center

Graduate Programs & Research

Advisor

Sanjay Modak

Advisor/Committee Member

Hammou Messatfa

Campus

RIT Dubai

Plan Codes

PROFST-MS

Download

Share

COinS