UNBIASED ESTIMATION OF THE AUTOCORRELATION COEFFICIENTS FOR A STATIONARY GAUSSIAN PROCESS WITH KNOWN MEAN AND VARIANCE USING THE SMOOTH CLIPPING FUNCTION

Abstract

This paper discusses the estimation of autocorrelation coefficients for a stationary Gaussian process with known mean and variance. An unbiased estimator which is an extension of the sample autocorrelation and the simplified estimator is proposed. An analytical expression for its variance is derived, and efficiency of the estimator is studied under two models.