A Comparative Study on Semiparametric Estimation in Partially Linear Single-Index Model

• Activity Number: 372
• Type: Contributed
• Date/Time: Tuesday, August 6, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Korean International Statistical Society
• Author(s): Young-Ju Kim*+
• Companies: Kangwon National University
• Keywords: Boosting ; partial spline ; partially linear model ; project pursuit regression ; single-index

Abstract

Partially linear single-index models are regarded as generalized versions of linear regression that allow the dimension reduction of multivariate predictors into a univariate index, while other predictors remain linearly associated with the response. We propose a semiparametric method based on partial splines for estimating the unknown function and regression parameters. Three methods - project pursuit regression, average derivative method, and boosting - are considered for estimating the single-index parameter. The relative performances of the methods are illustrated by simulation study and a real-world data example.