Title: Censored autoregressive regression modeling using the R package ARCensReg
Authors: Katherine Andreina Loor Valeriano - University of Campinas (Brazil)
Victor Hugo Lachos Davila - University of Connecticut (United States)
Larissa Avila Matos - Campinas State University (Brazil)
Christian Eduardo Galarza Morales - Escuela Superior Politecnica del Litoral (Ecuador)
Fernanda Schumacher - The Ohio State University (United States) [presenting]
Abstract: In several applications, data are collected over time and may contain censored or missing observations, making it impossible to use standard statistical procedures. The analysis of censored linear regression models with autoregressive errors is discussed using the R package ARCensReg, which implements maximum likelihood estimation via a stochastic approximation of the EM algorithm. The package was recently updated and accounts for both normal and Student-$t$ distributions, the latter distribution being particularly relevant for dealing with data that contain outlier observations. The use of the package for model selection and estimation will be illustrated using a real data set.