Information Compression, Structure Analysis, and Decision Making with a Correlation Matrix.
Abstract
Correlation functions have spawned a wide variety of data processing methods in a number of scientific disciplines. Under the broad umbrella of pattern recognition methodology, this report summarizes and coordinates several digital computer algorithms, all based on correlation matrices, for reducing the size of a data matrix and for producing algorithms which classify patterns. Each row of the data matrix represents a pattern and each column represents a feature. Correlation-based cluster-seeking algorithms which provide information used in choosing a decision-making strategy are also discussed. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1970
- Accession Number
- AD0720811
Entities
People
- Richard C. Dubes
Organizations
- Michigan State University