A Statistical Decision Theory for Non-Stationary Systems.
Abstract
A modification of the statistical theory of hypotheses testing that should be of interest in connection with sonar detection problems is presented. The problem can be described as one of making hypotheses about a physical system that is changing during the observation process (non-stationary system). When the states of the system follow a Markov law, a set of functions forming a sufficient statistic can be determined recursively; from them, it is possible to determine the optimum test according to any criterion. The procedure is easy to implement and provides - at least theoretically - an optimum way of taking the past information into account. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- May 01, 1971
- Accession Number
- AD0723868
Entities
People
- Christian Fabry
Organizations
- SACLANT ASW Research Centre