Background: With the increasing complexity of urban societies and the diversity of criminal behaviors, crime prediction has become an important area of criminology and security policy studies. Data-driven technologies and machine learning methods enable the analysis of large volumes of criminal, social, and demographic data and help identify behavioral patterns of criminals. These approaches can predict the probability of crime by combining personal information, criminal history, and environmental characteristics and provide the basis for designing targeted preventive programs. In addition, analyzing complex data and extracting hidden patterns allows the identification of crime hotspots, optimizing the allocation of law enforcement resources, and increasing the efficiency of crime prevention policies. For this reason, integrating machine learning approaches into criminology studies is a new and necessary approach to reduce criminal activity and promote social security.
Objective: The purpose of this study is to investigate the application of machine learning techniques to analyze criminal activity data and propose data-driven approaches to crime prediction. In fact, this study will focus on identifying the most effective machine learning models for predicting criminal behavior, understanding the key factors affecting crime rates, and offering practical insights for law enforcement agencies. By doing so, this study seeks to help define more effective and proactive crime prediction strategies and ultimately increase public safety and reduce the incidence of criminal behavior.
Methodology: In terms of the type of implementation and purpose of the study, it is a descriptive-survey type. In this study, the population and statistical sample of this study included information on 3,000 individuals who have a criminal record, which is in the form of 8 qualitative variables, including age, gender, nationality, education, marital status, type of crime, type of conviction, an