Predicting Crime Time Intervals Using Machine Learning Models

Understanding the time interval of crime can help optimize patrols and guards to identify crime-prone areas and estimate the time prone to crime. The urgency of this research lies in the need to develop more efficient methods for analyzing and preventing crime. By understanding the time pattern of crime, law enforcement can improve more effective prevention and law enforcement strategies. The methods used are DT, XGBoost, and CatBoost. This method was chosen because of its superior ability to handle large, complex, and unbalanced datasets. The evaluation was carried out using MAPE to measure the level of accuracy of crime clock predictions. The results show that XGBoost successfully predicts the time pattern of crimes with a MAPE of 8.29%, indicating a high level of accuracy. These results can be effectively applied to predict time-based crimes, helping to make better preventive decisions and improving the efficiency of security resource allocation.