Sir i think if query match ontology exist in database. The query should match many ontology but to what extend it match for the particular is term as good ontology. Means if ontology fulfill user's query with exact ontology and finally return single answer for query.

From the tutorial "Ontology-Driven Conceptual Modelling" by Nicola Guarino during ER2002, a better ontology is one where the possible models expressed by the ontology matches closer to the intended models to be expressed. A bad ontology would includes a lot of unintended models.

It depends on what you mean by "good" and "better" ontology.

There is the structure and there are the semantics.

So one has to consider both aspects in the process of evaluating an ontology.

There are numerous ontology evaluation criteria, here are some of the ones that I personally find interesting:

http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2013_Communique

http://www.sciencedirect.com/science/article/pii/S0169023X0400223X

Ontology Maturing is a continuous process that starts with new emergence of idea (which) can be represented through a new tag or relation within the ontology to be then more discussed by the community and/or refined, then shared as a common vocabulary which then can be part of a light-weight formal ontology (Mature ontology).

in the MATURE project (http://mature-ip.eu/) you can find some articles for knowledge maturing and ontology maturing models such as: (Ontology Maturing:

a Collaborative Web 2.0 Approach to Ontology Engineering)

http://www.researchgate.net/publication/221022560_Ontology_Maturing_a_Collaborative_Web_2.0_Approach_to_Ontology_Engineering/file/e0b49527d498416f14.pdf

the research of this is still going on and we are in Learning layers project (http://learning-layers.eu/) taking this research further to tackle the problem of scaling up technology at the work place.

Conference Paper Ontology Maturing: a Collaborative Web 2.0 Approach to Ontol...

Read this Article:

"The Ontology Maturing Approach for Collaborative and

Work Integrated Ontology Development: Evaluation

Results and Future Directions"

form the same research group, here are some indicators:

...............the quality of an ontology within the context of a semantic

application as a balance of three different aspects:

a) Appropriateness. An ontology needs to be an appropriate representation of the

domain with respect to the purpose of the ontologies required for a semantic application

so that it is actually useful. This is only possible when we have a tight

coupling and immediate mutual feedback between changes to the ontology and use

of its elements, e.g., for search or annotations. That means, we need a quick, simple

and work-integrated way to adapt and modify the ontologies.

b) Social Agreement. An ontology needs to represent a shared understanding among

all stakeholders. Thus, successful ontology construction is a social and collaborative

learning process within the communities of its users. The involved individuals

deepen by and by their understanding of the real world and of an (appropriate)

vocabulary to describe it.

c) Formality. The formalization of ontologies is not possible completely from scratch.

In particular for emerging ideas and concepts, it is not possible to directly integrate

6 International Workshop on Emergent Semantics and Ontology Evolution

them into an ontology as they are not clearly defined, yet. That means, the development

of an ontology underlies a process of continuous evolution where different

levels of formality might co-exist within one ontology. The outcome is an adequate

level of formality in the ontology, avoiding both over formalization and the inability

to apply semantic algorithms.

We did a little bit of work looking at how active ontologies were and how they changed in the biomedical domain. We attempted to extract some levels of 'maturity' similar to those in the software capability maturity model , but only vaguely and largely in the context of development only. Paper is here if of interest:

http://www.ncbi.nlm.nih.gov/pubmed/22554701

Domain ontology maturity goes by the growth of the domain. Ontology should yield high precision and high recall. A low recall is an indicator for the need of further development, while low precision indicates further refinement.

“All models are wrong, but some are useful.” George E. P. Box

http://ontology.buffalo.edu/smith/articles/EvaluationOfOntologies.pdf

In our Lab, NEMO, nemo.inf.ufes.br, we commit to the notion that an ontology is, in the first place, a reference conceptual model. We also commit to a representation language OntoUML, which is designed using Unified Foundation Ontology - UFO. In NEMO there are a number of tools, for modeling, validating, transforming conceptual models onto OWL models. These transformations accounts for relevant design decisions concerning expressiveness, once modal characteristics of OntoUML cannot be taken into account in the target implementation platform, and OCL rules transformation.

In this scenario, we have not considered a maturity model, although ontology quality improving approaches are a relevant research topic within the Lab.

For this latter topic, consider visiting the NEMO page and review the works of Giancarlo Guizzardi with Tiago Sales and Veruska Zamborlini, as well as, Bernardo Braga, and Alessander Benevides, to cite some.

Please look at a survey on approaches to ontology evaluation:

Samir Tartir, I. Budak Arpinar, Amit P. Sheth

Ontological Evaluation and Validation

https://www.researchgate.net/publication/226387227_Ontological_Evaluation_and_Validation?ev=srch_pub

Chapter Ontological Evaluation and Validation

To my knowledge, one of best ways to evaluate an ontology would be applying it in one practical software or application. I made an ontology over learning objects and used it to expose the learning materials as Linked Data format. In this way, when I used my ontology in an actual software, I realized that how a GOOD ontology can facilitate not only the programming stuff but also understanding the logics, concepts and classes in machine level.

I agree with Mohamed Al Smadi above.

In the development of the CIDOC CRM (ISO21127) and its extensions we apply the following principles:

A) The ontology is only valid with respect to a certain information function or form of discourse, such as "understanding particular facts about discrete things and events from the past at a human scale". (This would be different from astronomy, weather forecasts, car manufacturing etc.) Undefined or overly wide functionality leads to ambiguities in concepts and fuzzy boundaries, which is counterproductive.

B) Ultimately the ontology should provide the constant terms, in particular relationship terms, in order to structure information in a way that fits with the kinds of research questions or information requests of the users. We analyze explicit primary research questions to validate and mature the ontology. The more research questions and their variants can be answered using a set of concepts from an ontology, the more mature and effective the ontology is. The smaller the ontology is, the better is can be maintained.

The adequate generalization of particular research questions is a knowledge engineering task in its own right and requires domain knowledge.

C) Classes in an ontology are the carriers of properties. Only propositions represent knowledge ("I know that"), and propositions are based on properties. In order to have well defined properties, classes need to be well defined. We require a class definition to comprise the following: a) What is the substance of an instance? (letters, paper, process, etc.?). b) What is the identity condition that allows to distinguish two instances of this class (e.g., being composed of different letters?) c) What are the rules to decide unity for this class, i.e., if something is part or portion of an instance ? d) by which event or process an instance comes into existence and ends to exist. (e.g., by writing :: by loosing all copies)

Only if these four conditions are clarified, any described instance in an information system is unambiguous with respect to other instances and its behavior in the system over time. These maturity criteria come not by themselves out of unguided practice. They are again outcome of a knowledge engineering process.

Unfortunately most ontologies make definitions that cannot be verified against a domain.

if an ontology is conceived to be a conceptual model then it can use a ISO/IEC or CMI maturity model, otherwise, it depends on their application. for example, I have worked on the a validation and verification model since 2005 "OWLtoAspectj", "OWLtoUML", and "OWLtoJava". Thus I did my maturity model according to my development process.

For the question on detecting whether your ontology needs to be further developed or not, I can recommend my work on ontology completion. The purpose of ontology completion is, with the help of a domain expert to find out whether there are any missing axioms or individuals in the ontology. Here, "missing" is in the sense that an axiom or individual exists in the application domain, but is not represented in the ontology. The method guarantees to ask the minimum number of questions to a domain expert and complete the missing knowledge. You can have a look at the paper in the attachement.

Conference Paper Completing Description Logic Knowledge Bases Using Formal Co...