Analysis of Raw Data
Arthur David Snider · 2017
This chapter shows how to efficiently estimate the moments of a random process from a record of the process—a “run of data” or time series. It assumes that the data have been uniformly sampled and thus treat the random process as discrete. The chapter describes the concepts of stationarity and ergodicity and explains the limit concept in random processes. It discusses a sophisticated, extensively studied methodology for estimating autocorrelations of ergodic random processes using a finite record of data drawn from one specific realization of the process. The chapter also shows how the spectral estimation of autocorrelation is performed using the Bartlett’s method. It provides a discussion on the spectral analysis for continuous random processes.