Selection of Mixture Quantile Regression Models
The key issue of applying a conditional quantile model in applications is how to choose an appropriate quantile regression model. This dissertation proposes a series of new model selection approaches via penalized loss function plus a shrinkage operator to select an appropriate quantile regression m...
Uloženo v:
| Hlavní autor: | |
|---|---|
| Médium: | Dissertation |
| Jazyk: | angličtina |
| Vydáno: |
ProQuest Dissertations & Theses
01.01.2024
|
| Témata: | |
| ISBN: | 9798384016762 |
| On-line přístup: | Získat plný text |
| Tagy: |
Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
|
| Shrnutí: | The key issue of applying a conditional quantile model in applications is how to choose an appropriate quantile regression model. This dissertation proposes a series of new model selection approaches via penalized loss function plus a shrinkage operator to select an appropriate quantile regression model for a real application so that the proposed method not only selects an appropriate conditional quantile but also estimates the related parameters simultaneously. In Chapter 1, I summarize recent development in the literature of dynamic quantile regression models, the motivation and framework of this paper. Then in Chapter 2, I establish the asymptotic properties of the proposed estimator, including the rate of convergence and asymptotic normality and abnormality. Particularly, when the true coefficient parameters may be on the boundary of the parameter space and the marginal parameters are in an unidentified subset of the parameter space, I show that the limiting distribution for boundary parameter estimator is half-normal and the estimator for unidentified parameter converges to an arbitrary value. Monte Carlo simulation studies are conducted for demonstrating the finite sample performance of the proposed approach and the proposed method is used to investigate the stock return.Dynamic changes often occurs in scientific areas, such as economics and finance, due to changes in technologies, policies, international relations, preferences, etc. In order to capture the dynamic structure of real world data, I then extend the model in Chapter 2 into dynamic context where the models might change over time or different levels of another variable. Therefore, in Chapter 3 and 4, the weights and marginal parameters are assumed to be covariate-dependent or time-varying, respectively, and the corresponding semiparametric estimation and selection methods are proposed. Similarly, empirical studies are conducted for the same purpose as in Chapter 2. |
|---|---|
| Bibliografie: | SourceType-Dissertations & Theses-1 ObjectType-Dissertation/Thesis-1 content type line 12 |
| ISBN: | 9798384016762 |