Modeling the risks of the confession process of the accused of criminal offenses based on survival concept

Abstract

Based on statistical survival analysis, the assessment and forecasting of the risks of pleading guilty to criminal offenses in conditions of incomplete data are carried out. Risk function is constructed to estimate the probability of confession of suspects at certain stages (time periods) of the trial. The Kaplan-Meier model is applied to calculate the chances of obtaining confession evidence after the end of the trial in criminal proceedings. Differences in the decision to admit guilt for two groups of defendants: in the commission of a criminal offense by one person and a group of persons are investigated. Cox regression model is constructed to establish the interconnection between the stages of the pre-trial investigation, at which the accused gives confessions, with the duration of the investigation and the method of prosecution.