Predicting Reasoner Performance on ABox Intensive OWL 2 EL Ontologies

Abstract

Reasoner performance prediction for ontologies in the OWL 2 language has been studied so far from different dimensions. One key aspect of these studies has been the prediction of how much time a particular reasoning task for a given ontology will consume. Several approaches have adopted machinelearning techniques to predict time consumption of different reasoning tasks depending on features of the input ontologies. However, these studies have focused on capturing general aspects of the ontologies (i.e., mainly the complexity of their TBoxes), while paying little attention to ABox details. ABox information is particularly important in real-world scenarios, where data volumes are much larger than data-describing schema information. In this paper, we introduce the notion of ABox intensity in the context of predicting reasoner performance and to improve the representativeness of ontology metrics by developing new metrics that focus on ABox features of OWL 2 EL ontologies. Our experiments show that taking into account the intensity through our proposed metrics contributes to overall prediction accuracy for ABox intensive ontologies.

Understanding and improving ontology reasoning efficiency through learning and ranking

“…We plan to extend our methodology to support ABox reasoning, and investigate support of other non-standard reasoning problems. Continuing on our work on ABox-intensive EL ontologies [38] and reasoning on the Android platform [39], we will further investigate sources of ABox reasoning hardness by studying structural and syntactic properties of ABoxes. Performance prediction and optimisation utilising machine learning techniques is particularly interesting and relevant in the context of ontology-based data access (OBDA) [5], where a large database is enhanced by an ontology, and (conjunctive) query answering on the database requires ontology reasoning [61].…”

“…In a similar spirit, we investigated the prediction of reasoning time of ABox-intensive OWL 2 EL ontologies [38] and energy consumption of reasoning tasks on the Android platform [39].…”

A machine learning approach for optimizing heuristic decision‐making in Web Ontology Language reasoners

“…In addition to the ones mentioned in Table 19, we also added as features the occurrence number of different types of OWL object properties, including functional, transitive, symmetric, inverse functional object properties. 11 Another feature that we found to be useful in our case due to the importance of ABox individuals, is the average population, which indicates the ratio of concept instances to the number of classes in an ontology. 43 Obtaining some of the features such as the depth of the class hierarchy can be very expensive, especially for big-size ontologies.…”

“…ML aids with effective and successful decision-making in different states related to reasoning about ontologies, for example, assigning the most optimal OWL reasoner to a specific ontology based on the ontology's features. 11 Furthermore, ontology features play a pivotal role in predicting reasoner performance. Many papers have revealed the impact of ontology features on ML approaches for reasoners.…”

Learning the Right Expansion-ordering Heuristics for Satisfiability Testing in OWL Reasoners

“…Machine learning aids with effective and successful decision-making in different non-deterministic states related to reasoning about ontologies, e.g., assigning the most optimal DL reasoner to a specific ontology based on the ontology's features [19]. Ontology features play a pivotal role in predicting reasoner performance.…”