A Data-Driven Forecasting Model for Active Offenders on Electronic Monitoring Systems in Turkiye

The electronic monitoring of offenders is an increasingly popular technique in the criminal justice system. Worldwide, these systems are effectively utilized to monitor individuals on probation as they serve their sentence within the community. The use and significance of electronic monitoring systems are increasing day by day in Turkiye. This paper presents a complete ensemble empirical mode decomposition with adaptive noise and kernel based meta-extreme learning machine hybrid forecasting model using data on active offenders convicted of different crimes between 2013 and 2021 in Turkiye. Thanks to the proposed model, it is aimed to plan the equipment that will be needed and to provide optimal system management by observing the development of electronic monitoring systems in Turkiye. To validate the proposed model, it is compared with some state- of-the art model. The superiority of the proposed model is shown using some performance metrics. Moreover, the current status of electronic monitoring systems in Turkiye from the past to the present is shown statistically. While most studies on electronic monitoring focus on its financial or legal dimension, this paper uses a data-driven forecasting approach for optimal planning.