Accession Number : AD0757602

Title :   The Analyses of Stationary Stochastic Processes Using Spectral and Autoregression Techniques. Volume I.

Descriptive Note : Technical rept.,

Corporate Author : ARMY MATERIEL SYSTEMS ANALYSIS AGENCY ABERDEEN PROVING GROUND MD

Personal Author(s) : Groff,John N.

Report Date : SEP 1972

Pagination or Media Count : 122

Abstract : The report deals with the time and frequency domain analyses of stationary stochastic processes, i.e., a process whose statistics are time invariant over some interval of time. Algorithms and software have been developed which permit such analyses given either an ensemble of time histories which are representative of the process or merely a single time history of the process. The latter case is of particular interest since usually only limited data is available for such analyses. The report shows that if a process is gauss Markov in addition to being stationary then a single time history of sufficient length is adequate to perform frequency domain analysis or more precisely spectral analysis. The report also demonstrates the applicability of this methodology as an alternate approach to Monte Carlo simulation when suitable assumptions are satisfied. Such an application when feasible could result in a sizeable reduction in Monte Carlo budget requirements. (Author Modified Abstract)

Descriptors :   (*INFORMATION THEORY, TIME SERIES ANALYSIS), (*STOCHASTIC PROCESSES, SPECTRUM ANALYZERS), POWER SPECTRA, PROBABILITY DENSITY FUNCTIONS, RANDOM VARIABLES, REGRESSION ANALYSIS, GRAPHICS, COMPUTER PROGRAMS, TRACKING

Subject Categories : Statistics and Probability
      Cybernetics

Distribution Statement : APPROVED FOR PUBLIC RELEASE