Designing Quality Requirements, Metrics and Indicators for Core Ontologies: Results of a Comparative Study for Process Core Ontologies

TL;DR

This paper develops a structured approach for defining and evaluating quality requirements, metrics, and indicators for core ontologies, demonstrated through a comparative study of process core ontologies.

Contribution

It introduces a comprehensive quality model with metrics and indicators for core ontologies, and applies it to evaluate and improve process core ontologies.

Findings

Metrics and indicators effectively evaluate ontology quality.

Comparison results highlight strengths and weaknesses of the evaluated ontologies.

Improvement strategies enhance ontology quality based on evaluation outcomes.

Abstract

This preprint specifies quality requirements for a core ontology whose ontological elements such as terms, non-taxonomic relationships, among others, are based on a foundational ontology. The quality requirements are represented in a quality model that is structured in the form of a requirements tree composed of characteristics and attributes to be measured and evaluated. An attribute represents an atomic aspect of an entity, that is, an elementary non-functional requirement that can be measured by a direct or indirect metric and evaluated by an elementary indicator. In contrast, characteristics that model less atomic aspects of an entity cannot be measured by metrics, but rather are evaluated by derived indicators generally modeled by an aggregation function. Therefore, this preprint shows the design of direct and indirect metrics in addition to the design of elementary indicators, which are used to implement measurement and evaluation activities to obtain the results of a quality requirements tree. In particular, this document shows the applicability of the designed metrics and indicators that are used by a evaluation and comparison strategy. Two process core ontologies were preselected, evaluated and compared in order to adopt strengths in the target entity named ProcessCO. The data and information resulting from this study are also recorded, as well as the outcomes of the revaluation after improvement of the target entity.