PATTERN RECOGNITION OF STOCHASTIC PROCESSES.
Abstract
The class of differential processes are considered. For their class of processes, it is shown that there is an advantage to transforming the observed process x(t) into an infinite vector. Having derived the properties of the vector, it is then shown that the Decision Theory solution to the K-category recognition problems can be formulated in terms of the components of this vector. An example of signal detection is then worked out. It yields the optimum detector when the signal is a saw-tooth function. Finally, the K-category solution is explicitly derived.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jul 07, 1966
- Accession Number
- AD0489190
Entities
People
- Joel Owen