Adaptive Neuro-Fuzzy Model for Vehicle Theft Prediction and Recovery

Abstract

AbstractVehicle theft is continuously being reported as a global prevalent crime and the traditional mode of combating vehicle theft is faced with abnormalities hindering accurate, timely prediction and recovery of stolen vehicles from criminals. In this paper, we use Adaptive Neuro-Fuzzy Inference System (ANFIS) - a computational Artificial Intelligence (AI) technique to develop a model for minimizing investigation time and the number of deployed security operatives towards achieving a high success rate in the prediction, detection and recovery of stolen vehicles. A collection of vehicle theft and recovery data for (6) six consecutive years with fourteen (14) attributes collated by the Criminal Investigation Department of the Nigeria Police Force, Abeokuta, Ogun state were further analyzed through Dimensionality Reduction technique and Routine Activity Theory (RAT) approach to extract the most significant features. Datasets were sub-divided into 60%, 20% and 20% for training, testing and validating the model respectively. A significant result of 92.91% obtained with the Adaptive Neuro-Fuzzy Inference System (ANFIS) model showed that it is most efficient in predicting, detecting and recovering stolen vehicles as compared with other machine learning algorithms such as Random Tree, Naïve Bayes, J48 and Decision Rule of prediction accuracies of 86.51%, 71.24%, 67.68% and 55.73% respectively. KeywordMachine learningNeuro-fuzzyPredictionRecoverySelectionSignificant featuresVehicle theft