Paper
A Theory of Heuristic Information
- Authors:
- Chun-Hung Tzeng
- Abstract
- This paper introduces, based on probability, a general theory of a heuristic search and heuristic information in handling uncertainty. A search model consists of a probability space and a random variable. The task of the model is to search for heuristic information in order to approximate the random variable. Such heuristic information forms a Borel subfield of the probability space and the evaluation is the conditional expectation of the random variable relative to the Borel subfield. The paper considers comparisons and combinations of heuristic searches and studies their properties. As an application, this paper develops a pattern-recognition model, where the classical Bayes and Neyman-Pearson theorems are generalized. Similarity is applied to a pattern recognition, where a minimal representative system is introduced as a data clustering. To search for heuristic information is to compare an object with the representatives. The rough set approach in handling uncertainty is a special search model studied in this paper.
- Keywords
- Heuristic Information; Heuristic Search; Pattern Recognition; Similarity; Rough Sets
- StartPage
- 87
- EndPage
- 96
- Doi