Self-descriptive IF THEN rules from signal measurements: A holonic-based computational technique

Abstract

A holon is a bio-inspired conceptual entity that, like cells in a living organism, behaves as a part and a whole at the same time. Holonic systems have been the subject of intense research in the latest years due to their properties such as self-organization, self-similarity and capability of handling hierarchically-nested granularity levels. Lesser attention indeed has been paid by engineers to the aspect of self-description, i. e. the ability to describe itself in terms of self-contained descriptors. Self-description can be useful in measurement settings where the only available knowledge is embedded in data in terms of hidden rules behind observed signals. In this work, a heuristic technique is employed to extract self-descriptive IF THEN rules from measurement signals. These rules are considered holonic in that they represent a whole described in terms of relationships among their parts. An example taken from a real measurement scenario is reported and commented in detail.