The Crime Prediction of Criminal Activity Based on Weather Changes Towards Quality of Life

Abstract

Crime is a significant problem in society, and crime prevention is crucial. Factors such as politics, economics, culture, education, demographics, and employment have been identified as contributing to crime. Recent studies have also explored the relationship between weather and crime. Therefore, this research aims to identify the best-performing machine learning algorithm based on weather in Malaysia, using crime data from the Royal Malaysia Police and Meteorological Department from 2011 to 2020. Five machine learning algorithms were utilized, and the results showed that all algorithms had good prediction accuracy, with Gradient Boosted Trees performing the best, with an error rate of less than 23%. Location was found to be the most important feature in all the models. This study provides a valuable fundamental framework for environmental crime and social impact research scholars to conduct a more in-depth analysis of the prediction models. This study establishes a fundamental framework for scholars in environmental crime and social impact research to conduct in-depth analysis using prediction models, thereby contributing to a better understanding of the complex relationship between weather and crime, and aiding in the development of effective crime prevention strategies.