Abstract
Abstract : Single index model (SIMo) has become one of the most important semiparametric to overcome the
dimensionality problem. This model work on summarizes the effects of the independent variables within a single
variable and refers to them as the index. In this paper, the Bayesian hierarchical model is constructed to estimate the
parameters and the unknown nonparametric function for the single index when the response variable is censored. In
addition, to get a head start on finding the unknown nonparametric link function, we assume it follows the Gaussian
process distribution. The Laplace distribution will be used as a prior distribution for the coefficient vector β when
variables are being selected. The performance of the suggested technique is evaluated by applying it to simulation
examples and actual data.
dimensionality problem. This model work on summarizes the effects of the independent variables within a single
variable and refers to them as the index. In this paper, the Bayesian hierarchical model is constructed to estimate the
parameters and the unknown nonparametric function for the single index when the response variable is censored. In
addition, to get a head start on finding the unknown nonparametric link function, we assume it follows the Gaussian
process distribution. The Laplace distribution will be used as a prior distribution for the coefficient vector β when
variables are being selected. The performance of the suggested technique is evaluated by applying it to simulation
examples and actual data.
Keywords
Gaussian process distribution
Keywords: Bayesian Inference
Markov chain Monte Carlo methods; Variable selection.
Tobit regression