City Research Online

Copula Regression Spline Sample Selection Models: The R Package SemiParSampleSel

Wojtys, M., Marra, G. and Radice, R. ORCID: 0000-0002-6316-3961 (2016). Copula Regression Spline Sample Selection Models: The R Package SemiParSampleSel. Journal of Statistical Software, 71(6), doi: 10.18637/jss.v071.i06

Abstract

Sample selection models deal with the situation in which an outcome of interest is observed for a restricted non-randomly selected sample of the population. The estimation of these models is based on a binary equation, which describes the selection process, and an outcome equation, which is used to examine the substantive question of interest. Classic sample selection models assume a priori that continuous covariates have a linear or pre-specified non-linear relationship to the outcome, and that the distribution linking the two equations is bivariate normal. We introduce the R package SemiParSampleSel which implements copula regression spline sample selection models. The proposed implementation can deal with non-random sample selection, non-linear covariate-response relationships, and non-normal bivariate distributions between the model equations. We provide details of the model and algorithm and describe the implementation in SemiParSampleSel. The package is illustrated using simulated and real data examples.

Publication Type: Article
Additional Information: This is an open access article licensed under a Creqtive Commons Attribution 3.0 Unported License.
Publisher Keywords: copula, non-random sample selection, penalized regressi on spline, selection bias, R
Subjects: H Social Sciences > HA Statistics
Q Science > QA Mathematics
Departments: Cass Business School > Actuarial Science & Insurance
URI: http://openaccess.city.ac.uk/id/eprint/20926
[img]
Preview
Text - Published Version
Available under License Creative Commons: Attribution 3.0.

Download (690kB) | Preview

Export

Downloads

Downloads per month over past year

View more statistics

Actions (login required)

Admin Login Admin Login