Semiparametric and nonparametric estimation for dynamic quantile regression models

Since quantile regression was proposed by Koenker and Bassett (1978), recently, it has been successfully applied to various applied fields such as finance and economics as well as biology. In this dissertation, I consider two classes of quantile regression models for dynamic time series data: nonparametric and semiparametric quantile regression models with a functional or partially functional coefficient structure.;Firstly, I develop an estimate procedure to estimate functional coefficients by using local linear approximations under dynamic time series data. I derive the local Bahadur representation of the local linear estimator under alpha-mixing conditions and establish the consistency and the asymptotic normality of the estimator.;Secondly, I derive the n -consistency estimator for parameters in semi-parameterc model by using average method for beta-mixing time series. Also I establish the consistency and the asymptotic normality of the proposed estimator.;The programming involved in the proposed estimation procedures is relatively simple and it can be modified with few efforts from the existing programs for the linear quantile model. A comparison of the local linear quantile estimator with other methods is presented. Simulation studies are carried out to illustrate the performance of the estimates. An empirical application of the model to the exchange rate time series data and the well-known Boston house price data further demonstrates the potential of the proposed modeling procedures.