Accession Number:

ADA285472

Title:

Pattern Theoretic Knowledge Discovery

Descriptive Note:

Final rept. Dec 1993-Aug 1994

Corporate Author:

WRIGHT LAB WRIGHT-PATTERSON AFB OH

Personal Author(s):

Report Date:

1994-08-01

Pagination or Media Count:

53.0

Abstract:

Among the future research directions of Knowledge Discovery in Databases is the ability to extract an overlying concept relating data objects that are useful to the investigator. Some of the current limitations involve the search complexity and what it means to be useful. The Pattern Theory research crosses over in a natural way to the aforementioned domain. The goal of this paper is threefold. First, we wish to present a new approach to the problem of learning by Discovery and robust pattern finding in general. Second, we will show its performance by exhibiting several learning curves. Third, from a practical standpoint, we wish to explore the current limitations of a Pattern Theoretic Discovery and Databases problem. Function decomposition is the central core of Pattern Theory. The development allows us to discuss the notion of patterns, and thus, the notion of useful, in a formal manner. Pattern Theory, Function Decomposition, Machine Learning Patterns, Knowledge Discovery

Subject Categories:

  • Cybernetics

Distribution Statement:

APPROVED FOR PUBLIC RELEASE