Bayesian quantile regression for semiparametric models
Quantile regression has recently received a great deal of attention in both theoretical and empirical research. It can uncover different structural relationships between covariates and responses at the upper or lower tails, which is sometimes of significant interest in econometrics, educational and...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Language: | English |
| Published: |
2013
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/54830 |
| _version_ | 1826113559114809344 |
|---|---|
| author | Hu, Yuao |
| author2 | Lian Heng |
| author_facet | Lian Heng Hu, Yuao |
| author_sort | Hu, Yuao |
| collection | NTU |
| description | Quantile regression has recently received a great deal of attention in both theoretical and empirical research. It can uncover different structural relationships between covariates and responses at the upper or lower tails, which is sometimes of significant interest in econometrics, educational and medicine applications. The methodologies of quantile regression for linear models have been well developed in both frequentist and Bayesian contexts. However, there has been relatively less work focusing on quantile regression for nonparametric models or semiparametric models, especially from a Bayesian perspective.
The principal goal of this work is to propose efficient approaches to implement Bayesian quantile regression with two kinds of semiparametric modes, single-index models and partially linear additive models, using an asymmetric Laplace distribution which provides a mechanism for Bayesian inference of quantile regression. With carefully selected priors, we build hierarchical Bayesian models and design effective Markov chain Monte Carlo algorithms for posterior inference. We compare the proposed methods with some existing methods through simulation studies and real data applications. |
| first_indexed | 2024-10-01T03:25:12Z |
| format | Thesis |
| id | ntu-10356/54830 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T03:25:12Z |
| publishDate | 2013 |
| record_format | dspace |
| spelling | ntu-10356/548302023-02-28T23:36:12Z Bayesian quantile regression for semiparametric models Hu, Yuao Lian Heng School of Physical and Mathematical Sciences DRNTU::Science::Mathematics::Statistics Quantile regression has recently received a great deal of attention in both theoretical and empirical research. It can uncover different structural relationships between covariates and responses at the upper or lower tails, which is sometimes of significant interest in econometrics, educational and medicine applications. The methodologies of quantile regression for linear models have been well developed in both frequentist and Bayesian contexts. However, there has been relatively less work focusing on quantile regression for nonparametric models or semiparametric models, especially from a Bayesian perspective. The principal goal of this work is to propose efficient approaches to implement Bayesian quantile regression with two kinds of semiparametric modes, single-index models and partially linear additive models, using an asymmetric Laplace distribution which provides a mechanism for Bayesian inference of quantile regression. With carefully selected priors, we build hierarchical Bayesian models and design effective Markov chain Monte Carlo algorithms for posterior inference. We compare the proposed methods with some existing methods through simulation studies and real data applications. DOCTOR OF PHILOSOPHY (SPMS) 2013-08-30T03:49:08Z 2013-08-30T03:49:08Z 2013 2013 Thesis Hu, Y. (2013). Bayesian quantile regression for semiparametric models. Doctoral thesis, Nanyang Technological University, Singapore. https://hdl.handle.net/10356/54830 10.32657/10356/54830 en 123 p. application/pdf |
| spellingShingle | DRNTU::Science::Mathematics::Statistics Hu, Yuao Bayesian quantile regression for semiparametric models |
| title | Bayesian quantile regression for semiparametric models |
| title_full | Bayesian quantile regression for semiparametric models |
| title_fullStr | Bayesian quantile regression for semiparametric models |
| title_full_unstemmed | Bayesian quantile regression for semiparametric models |
| title_short | Bayesian quantile regression for semiparametric models |
| title_sort | bayesian quantile regression for semiparametric models |
| topic | DRNTU::Science::Mathematics::Statistics |
| url | https://hdl.handle.net/10356/54830 |
| work_keys_str_mv | AT huyuao bayesianquantileregressionforsemiparametricmodels |