SFB 303 Discussion Paper No. A-544

Author: Zhu-Yu Li , Zu-Di Lu, Gen-Xiang Chai
Title: Nonparametric Estimation of Density Function for a Sample Process under Mixing Condition
Abstract: Nonparametric density estimation is a useful tool for examining the structure of data, in particular, for the stationary time series, since usually it is really difficult to find the real marginal density of the series. Some papers contributed to this aspect for alpha-mixing stationary sequence can be found in the literature, e.g., Robinson (1983), Tran (1989, 1990). However, just as Tran et al (1996) stressed, yet there are a great number of processes which may not be alpha-mixing. In this paper, we will adopt a nonparametrical method to estimate unknown density function of a sample data process which is based on relaxing phi-mixing assumptions (see Billingsley (1968) and Bierens (1983)). Uniformly weak and strong consistency and the convergence rates of the estimator we adopted will be discussed, and some numerical examples will be given.
Keywords: Relaxing phi-Mixing Sample, Stationary Sample Process, Non- parametric Statistics, Kernel Density Estimators, Asymptotic Consistency, Rate of Convergence.
JEL-Classification-Number: C13, C14, C22.
Creation-Date: 1997
URL: ../1997/a/bonnsfa544.pdf"

SFB 303 Homepage