Accession Number : AD0878406

Title :   Computation of the Conditional Variance with Uncertain Observations.

Descriptive Note : Technical rept. Mar-May 70,

Corporate Author : JOHNS HOPKINS UNIV BALTIMORE MD CARLYLE BARTON LAB

Personal Author(s) : Yurtseven, H. O.

Report Date : DEC 1970

Pagination or Media Count : 33

Abstract : A scalar estimation problem with uncertain observations is considered. It is different from the classical estimation problem in the sense that the observation contains the signal to be estimated with a given probability and thus it may contain noise alone. Nahi has derived the general formulation where he assumed that the best estimate is a linear function of observations. In this work, mean square errors of both Nahi's linear estimate and conventional nonlinear estimate are derived for a scalar problem and computed results are given for some choice of parameters. (Author)

Descriptors :   (*INFORMATION THEORY, STATISTICAL ANALYSIS), DECISION THEORY, STOCHASTIC PROCESSES, SET THEORY, MATRICES(MATHEMATICS), PROBABILITY DENSITY FUNCTIONS, ERRORS.

Subject Categories : Cybernetics

Distribution Statement : APPROVED FOR PUBLIC RELEASE