Concepts of an Estimation System, an Adaptive System, and a Network of Adaptive Estimation Systems
- 1 January 1969
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Systems Science and Cybernetics
- Vol. 5 (1) , 79-85
- https://doi.org/10.1109/tssc.1969.300248
Abstract
A clear distinction is made between two concepts. One concept is called an estimation system and the other an adaptive system. Using these two concepts as building blocks and a Bayes minimum conditional risk approach, it is possible to give meaning to the concept of a network of adaptive estimation systems. Under the framework of the approach, unsupervised estimation (estimation without a teacher) is a classical statistical problem as is supervised estimation. Also, supervised estimation is a degenerate case of estimation and a special case of adapting. In general, adapting without a teacher is impossible.Keywords
This publication has 9 references indexed in Scilit:
- On a class of unsupervised estimation problemsIEEE Transactions on Information Theory, 1968
- Asymptotic probability of error using two decision-directed estimators for two unknown mean vectors (Corresp.)IEEE Transactions on Information Theory, 1968
- Iterative computation of a posteriori probability forM-ary nonsupervised adaptation (Corresp.)IEEE Transactions on Information Theory, 1966
- Nonsupervised sequential classification and recognition of patternsIEEE Transactions on Information Theory, 1966
- Automatic off-line multivariate data analysisPublished by Association for Computing Machinery (ACM) ,1966
- Nonsupervised adaptive signal detection and pattern recognitionInformation and Control, 1964
- The Empirical Bayes Approach to Statistical Decision ProblemsThe Annals of Mathematical Statistics, 1964
- Learning to recognize patterns in a random environmentIEEE Transactions on Information Theory, 1962
- Non-Parametric Empirical Bayes ProceduresThe Annals of Mathematical Statistics, 1957