Classification Tree Model for Determining Society Unsafety Factors Convicted

Abstract

The global rise in crime rates is a result of the ineffectiveness of the traditional punishment system and the need to develop new approaches to applying relevant punitive measures to criminals, taking into account their level of danger to society. This paper presents a classification Decision Trees model to identify significant factors influencing the objective level of danger a convicted person poses to society. Applied research was conducted based on real data (individual characteristics and criminal history records) of 2,052 convicts serving sentences in correctional facilities in Ukraine. It was found that the most significant predictors for determining the level of danger to society were the number of previous incarcerations and the number of suspended sentences. Assessing the level of societal danger posed by convicts is one of the key aspects of upholding the principles of fair justice. The obtained results can serve as informational support for judicial decisions, ensuring a balance between societal protection and successful offender reintegration.