Frequency Domain Experiments: Spectral Amplification, Input-Output Correlations, and Model Parameters.
Abstract
A frequency domain approach to identifying a polynomial response model for a system was presented in an earlier document. Frequency domain experiments can help identify important input factors (or parameters) and interactions in systems where one can change factor values during an experimental trial and observe the response at periodic intervals. An example of such a system is a simulation run on a computer. In this paper we consider the relationships between spectral amplification, input-output correlation, and factor values for such experiments. Conditions are also given for which the frequency domain approach is equivalent to ranking prospective terms in a response model by their correlation with the output. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 25, 1985
- Accession Number
- ADA161979
Entities
People
- Lee Schruben
- V. J. Cogliano
Organizations
- Cornell University