Statistical Estimation and Decision Techniques Applied to Time Series Analysis for Communication and Tracking Problems.
Final rept. 1 Aug 70-31 Jul 72,
JOHNS HOPKINS UNIV BALTIMORE MD DEPT OF STATISTICS
Pagination or Media Count:
During the period of the contract, contributions were made to the theory and methodology of statistical inference for dynamic models. New procedures for recursive dynamic linear estimation of a trajectory were obtained and theoretically justified such procedures also have application in communications and physiological monitoring technology. In addition, a new method for forming point estimators was suggested, and shown to have optimal large sample properties. This methodology, and the accompanying theory, have direct application for forming sequential fixed-diameter confidence regions. Still further, estimators and tests of hypotheses were constructed for the parameters of certain structured multivariate models arising in the contexts of human performance evaluation, geology, andor communication theory. Some distributional representations and approximations were derived which promise to be of use for development and evaluation of inference procedures both in multivariate analysis, and in time series analysis including the analysis of random fields. Author
- Statistics and Probability