Possibilistic testing of OWL axioms against RDF data

Abstract

International audienceWe develop the theory of a possibilistic framework for OWL 2 axiom testing against RDF datasets, as an alternative to statistics-based heuristics. The intuition behind it is to evaluate the credibility of OWL 2 axioms based on the evidence available in the form of a set of facts contained in a chosen RDF dataset. To achieve it, we first define the notions of development, content, support , confirmation and counterexample of an axiom. Then we use these notions to define the possibility and necessity of an axiom and its acceptance/rejection index combining both of them. Finally, we report a practical application of the proposed framework to test SubClassOf axioms against the DBpedia RDF dataset

Predicting the Possibilistic Score of OWL Axioms Through Support Vector Regression

“…2. In particular, we used as a reference the same settings as in [10], briefly summarized hereafter: 1 we considered m = 722 SubClassOf axioms involving atomic classes which were exactly scored against DBpedia, 2 as well as their negations, in a set A (thus a total of n = 2m = 1444 formulas), computing sim for each pair; we took the possibility of each formula φ i ∈ A previously computed using the heuristic described in [15], henceforth identified as a value µ i = Π(φ i ) of the "acceptability" of φ i as an axiom.…”

“…The possibilistic axiom score we wish to be able to predict was proposed in [15], to which the reader is referred to for the details. Given a candidate OWL 2 axiom φ, expressing a hypothesis about the relations holding among some entities of a domain, a degree of possibility Π(φ) and of necessity N (φ) for φ are computed based the evidence available contained in an RDF dataset K.…”

“…The most popular candidate axiom scoring heuristics proposed in the literature are based on statistical inference (see, e.g., [4]). An alternative axiom scoring heuristics based on a formalization in possibility theory of the notions of logical content of a theory and of falsification and complying with an open-world semantics has recently been proposed [15]. While empirical evidence has been found that such a possibilistic scoring heuristics may lead to more accurate ontologies, the heavy computational cost of the heuristics makes it hard to apply in practice, unless some implementation tricks are devised (e.g., time capping).…”

“…In this work, we follow the same general scheme, but we apply support vector regression to obtain a simpler surrogate model and we compare our preliminary results to those obtained with the modified support vector clustering on the same dataset of SubClassOf (i.e., subsumption) axioms, whose possibilistic score has been previously determined by direct application of the heuristics on the DBpedia RDF dataset [15].…”

Converting OCL and CGMES Rules to SHACL in Smart Grids

“…In [10], instead of validating RDF data against a schema, a model is validated against RDF data. The purpose is validating model credibility or giving the capability of ontology learning.…”