Serial Pattern Acquisition: A Production System Approach
Abstract
A production system technique for recognizing regularities in serial patterns is presented in the context of the letter series extrapolation problem. The learning technique consists of creating an ordered set of production rules to represent the concept of a pattern, such that each rule is a hypothesis about which pattern contexts lead to which new pattern elements. The production system learning technique is compared with other series extrapolation methods and examples of series concepts learned by a computer implementation of the technique are given.
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 01, 1975
- Accession Number
- ADA013569
Entities
People
- D. A. Waterman
Organizations
- Carnegie Mellon University