Evaluation of ontologies

TLDR

Ontology evaluation is an emerging field lacking a deep core of preliminary ideas and guidelines. The paper summarizes prior ontology evaluation work and the criteria—consistency, completeness, conciseness, expandability, and sensitiveness—used to assess ontologies. The authors discuss error types in ontology taxonomies—circularity, exhaustive/nonexhaustive partition, redundancy, grammatical, semantic, incompleteness—and describe the evaluation process applied to the standard‑units ontology on the Ontology Server. © 2001 John Wiley & Sons, Inc.

Abstract

The evaluation of ontologies is an emerging field. At present, there is an absence of a deep core of preliminary ideas and guidelines for evaluating ontologies. This paper presents a brief summary of previous work done on evaluating ontologies and the criteria (consistency, completeness, conciseness, expandability, and sensitiveness) used to evaluate and to assess ontologies. It also addresses the possible types of errors made when domain knowledge is structured in taxonomies in an ontology and in knowledge bases: circularity errors, exhaustive and nonexhaustive class partition errors, redundancy errors, grammatical errors, semantic errors, and incompleteness errors. It also describes the process followed to evaluate the standard-units ontology already published at the Ontology Server. © 2001 John Wiley & Sons, Inc.