Prediction of Criminal Theft Locations at the Binjai Police Station using Historical Data and the KNN Algorithm

This research aims to predict the locations of theft crimes in the Binjai Police Department by utilizing historical data obtained from Binjai Police. The implementation of the K-Nearest Neighbor (KNN) algorithm was carried out to analyze crime patterns based on variables such as district location, time of occurrence, day of occurrence, and type of crime. The historical data were processed through normalization, splitting into training and testing datasets, and model evaluation using RapidMiner software. The results show that the KNN algorithm is able to classify with a fairly good level of accuracy, making it a useful basis for providing predictive information on theft-prone locations. These findings are expected to assist the police and the Binjai city government in formulating crime prevention strategies and increasing public awareness.