Stationary Time Series, Quantile Functions, Nonparametric Inference and Rank Transform Spectrum

摘要

In this dissertation, weak convergence results for dependent sequences are used to derive the asymptotic distribution of linear rank statistics for the two sample problem. It is shown that the asymptotic variance of linear rank statistics when computed from two independent time series depends on the spectrum of the rank transform time series. The behavior of the rank transform spectrum in terms of its relations to the original spectrum is also empirically examined. (Author)