Comparison of parametric and semi-parametric binary response models

Xiangjin Shen; Shiliang Li; Hiroki Tsurumi

A Bayesian semi-parametric estimation of the binary response model using Markov Chain Monte Carlo algorithms is proposed. The performances of the parametric and semi-parametric models are presented. The mean squared errors, receiver operating characteristic curve, and the marginal effect are used as the model selection criteria. Simulated data and Monte Carlo experiments show that unless the binary data is extremely unbalanced the semi-parametric and parametric models perform equally well. However, if the data is extremely unbalanced the maximum likelihood estimation does not converge whereas the Bayesian algorithms do. An application is also presented.

MoreLess

Year of publication:	2013
Authors:	Shen, Xiangjin ; Li, Shiliang ; Tsurumi, Hiroki
Publisher:	New Brunswick, NJ : Dep. of Economics, Rutgers, the State Univ. of New Jersey
Subject:	Semi-parametric binary response models \| Markov Chain Monte Carlo algorithms \| Kernel densities \| Optimal bandwidth \| Receiver operating characteristic curve \| Nichtparametrisches Verfahren \| Nonparametric statistics \| Markov-Kette \| Markov chain \| Monte-Carlo-Simulation \| Monte Carlo simulation \| Schätztheorie \| Estimation theory