Sequential Decision Procedures with Prespecified Error Probabilities and their Applications.
Abstract
A main problem in sequential analysis and pattern recognition is the design of sequential decision procedures in which it is possible to control the probability of error. A procedure is called optimum if it has a probability of error less than a specified value and minimizes the average observations cost among all procedures with probability of error less than this specified value. This research investigates first the existence of such optimum procedures and gives algorithms to obtain them.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 01, 1974
- Accession Number
- ADA006222
Entities
People
- Eric Persoon
- King Sun Fu
Organizations
- Purdue University