ANÁLISE DE SOBREVIVÊNCIA APLICADA A DADOS DE REINCIDÊNCIA AO CRIME

Main Article Content

Matheus Henrique Pimenta ZANON
http://orcid.org/0000-0003-0698-6360
Emílio Augusto Coelho BARROS

Abstract

The Brazilian prison system is a point of great discussion in our current political and social moment. The  last report (DEPEN, 2014) points out that Brazil is the fourth country with the largest number of individuals  deprived of their freedom in the world, esteemed although the rate of recidivism to crime for males is around 50% and for females is 35%. This work was based on the SECAT database information of 
the city of Primeiro de Maio-PR, Brazil, containing 356 individuals, covering the period from December 2009 to December 2015. Our event of interest is recidivism crime of each individual during the observation  period. To conduct this study, we are suggesting to use survival analysis techniques due to the fact that it contains censored observations. This paper advocate the use of distribution Burr XII with and without fraction of cure, besides the Weibull distribution with and without cure rate for modeling and analysing  the data, since there is evidence that many individuals have never faced the event of interest. To obtain the estimators of the proposed models, we have considered the frequentist and Bayesian inference approaches. 

Article Details

How to Cite
ZANON, M. H. P., & BARROS, E. A. C. (2018). ANÁLISE DE SOBREVIVÊNCIA APLICADA A DADOS DE REINCIDÊNCIA AO CRIME. Brazilian Journal of Biometrics, 36(3), 527–542. https://doi.org/10.28951/rbb.v36i3.228
Section
Articles