How Long Will It Take? Accurate Prediction of Ontology Reasoning Performance

Yong-Bin Kang, Jeff Z. Pan, Shonali Krishnaswamy, Wudhichart Sawangphol, Yuan-Fang Li

Research output: Chapter in Book/Report/Conference proceedingConference contribution

14 Citations (Scopus)

Abstract

For expressive ontology languages such as OWL 2 DL, classification is a computationally expensive task— 2NEXPTIME-complete in the worst case. Hence, it is highly desirable to be able to accurately estimate classification time, especially for large and complex ontologies. Recently, machine learning techniques have been successfully applied to predicting the reasoning hardness
category for a given (ontology, reasoner) pair. In this paper, we further develop predictive models to estimate actual classification time using regression techniques, with ontology metrics as features. Our largescale experiments on 6 state-of-the-art OWL 2 DL reasoners and more than 450 significantly diverse ontologies demonstrate that the prediction models achieve high accuracy, good generalizability and statistical significance. Such prediction models have a wide range of applications. We demonstrate how they can be used to efficiently and accurately identify performance hotspots in a large and complex ontology, an otherwise very time-consuming and resource-intensive task.
Original languageEnglish
Title of host publicationProceedings of the 28th Conference on Artificial Intelligence (AAAI 2014)
PublisherAAAI Press
Pages80-86
Number of pages7
Publication statusPublished - Aug 2014
Event28th AAAI Conference on Artificial Intelligence - Quebec Convention Center, Quebec, Canada
Duration: 27 Jul 201431 Jul 2014

Conference

Conference28th AAAI Conference on Artificial Intelligence
Abbreviated titleAAAI-14
CountryCanada
CityQuebec
Period27/07/1431/07/14

Fingerprint

Ontology
Learning systems
Experiments

Keywords

  • Ontology
  • Reasoning
  • Prediction
  • Performance hotspots

Cite this

Kang, Y-B., Pan, J. Z., Krishnaswamy, S., Sawangphol, W., & Li, Y-F. (2014). How Long Will It Take? Accurate Prediction of Ontology Reasoning Performance. In Proceedings of the 28th Conference on Artificial Intelligence (AAAI 2014) (pp. 80-86). AAAI Press.

How Long Will It Take? Accurate Prediction of Ontology Reasoning Performance. / Kang, Yong-Bin; Pan, Jeff Z.; Krishnaswamy, Shonali; Sawangphol, Wudhichart; Li, Yuan-Fang.

Proceedings of the 28th Conference on Artificial Intelligence (AAAI 2014). AAAI Press, 2014. p. 80-86.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Kang, Y-B, Pan, JZ, Krishnaswamy, S, Sawangphol, W & Li, Y-F 2014, How Long Will It Take? Accurate Prediction of Ontology Reasoning Performance. in Proceedings of the 28th Conference on Artificial Intelligence (AAAI 2014). AAAI Press, pp. 80-86, 28th AAAI Conference on Artificial Intelligence, Quebec, Canada, 27/07/14.
Kang Y-B, Pan JZ, Krishnaswamy S, Sawangphol W, Li Y-F. How Long Will It Take? Accurate Prediction of Ontology Reasoning Performance. In Proceedings of the 28th Conference on Artificial Intelligence (AAAI 2014). AAAI Press. 2014. p. 80-86
Kang, Yong-Bin ; Pan, Jeff Z. ; Krishnaswamy, Shonali ; Sawangphol, Wudhichart ; Li, Yuan-Fang. / How Long Will It Take? Accurate Prediction of Ontology Reasoning Performance. Proceedings of the 28th Conference on Artificial Intelligence (AAAI 2014). AAAI Press, 2014. pp. 80-86
@inproceedings{d7d9b91646f6424e828d6bee8b6fa45c,
title = "How Long Will It Take? Accurate Prediction of Ontology Reasoning Performance",
abstract = "For expressive ontology languages such as OWL 2 DL, classification is a computationally expensive task— 2NEXPTIME-complete in the worst case. Hence, it is highly desirable to be able to accurately estimate classification time, especially for large and complex ontologies. Recently, machine learning techniques have been successfully applied to predicting the reasoning hardnesscategory for a given (ontology, reasoner) pair. In this paper, we further develop predictive models to estimate actual classification time using regression techniques, with ontology metrics as features. Our largescale experiments on 6 state-of-the-art OWL 2 DL reasoners and more than 450 significantly diverse ontologies demonstrate that the prediction models achieve high accuracy, good generalizability and statistical significance. Such prediction models have a wide range of applications. We demonstrate how they can be used to efficiently and accurately identify performance hotspots in a large and complex ontology, an otherwise very time-consuming and resource-intensive task.",
keywords = "Ontology, Reasoning, Prediction, Performance hotspots",
author = "Yong-Bin Kang and Pan, {Jeff Z.} and Shonali Krishnaswamy and Wudhichart Sawangphol and Yuan-Fang Li",
note = "We would like to thank Rafael S. Gonc¸alves for providing us with the ontology and hotspots dataset. We are also grateful to Yuan Ren for the many helpful discussions.",
year = "2014",
month = "8",
language = "English",
pages = "80--86",
booktitle = "Proceedings of the 28th Conference on Artificial Intelligence (AAAI 2014)",
publisher = "AAAI Press",

}

TY - GEN

T1 - How Long Will It Take? Accurate Prediction of Ontology Reasoning Performance

AU - Kang, Yong-Bin

AU - Pan, Jeff Z.

AU - Krishnaswamy, Shonali

AU - Sawangphol, Wudhichart

AU - Li, Yuan-Fang

N1 - We would like to thank Rafael S. Gonc¸alves for providing us with the ontology and hotspots dataset. We are also grateful to Yuan Ren for the many helpful discussions.

PY - 2014/8

Y1 - 2014/8

N2 - For expressive ontology languages such as OWL 2 DL, classification is a computationally expensive task— 2NEXPTIME-complete in the worst case. Hence, it is highly desirable to be able to accurately estimate classification time, especially for large and complex ontologies. Recently, machine learning techniques have been successfully applied to predicting the reasoning hardnesscategory for a given (ontology, reasoner) pair. In this paper, we further develop predictive models to estimate actual classification time using regression techniques, with ontology metrics as features. Our largescale experiments on 6 state-of-the-art OWL 2 DL reasoners and more than 450 significantly diverse ontologies demonstrate that the prediction models achieve high accuracy, good generalizability and statistical significance. Such prediction models have a wide range of applications. We demonstrate how they can be used to efficiently and accurately identify performance hotspots in a large and complex ontology, an otherwise very time-consuming and resource-intensive task.

AB - For expressive ontology languages such as OWL 2 DL, classification is a computationally expensive task— 2NEXPTIME-complete in the worst case. Hence, it is highly desirable to be able to accurately estimate classification time, especially for large and complex ontologies. Recently, machine learning techniques have been successfully applied to predicting the reasoning hardnesscategory for a given (ontology, reasoner) pair. In this paper, we further develop predictive models to estimate actual classification time using regression techniques, with ontology metrics as features. Our largescale experiments on 6 state-of-the-art OWL 2 DL reasoners and more than 450 significantly diverse ontologies demonstrate that the prediction models achieve high accuracy, good generalizability and statistical significance. Such prediction models have a wide range of applications. We demonstrate how they can be used to efficiently and accurately identify performance hotspots in a large and complex ontology, an otherwise very time-consuming and resource-intensive task.

KW - Ontology

KW - Reasoning

KW - Prediction

KW - Performance hotspots

M3 - Conference contribution

SP - 80

EP - 86

BT - Proceedings of the 28th Conference on Artificial Intelligence (AAAI 2014)

PB - AAAI Press

ER -