Combining Ontology Class Expression Generation with Mathematical Modeling for Ontology Learning
Authors: Jedrzej Potoniec, Agnieszka Ławrynowicz
AAAI 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We did preliminary experiments for a few classes from DBpedia ontology, which generated interesting subhierarchies. They are available in the complementary materials4. |
| Researcher Affiliation | Academia | Jedrzej Potoniec and Agnieszka Ławrynowicz Institute of Computing Science, Poznan University of Technology ul. Piotrowo 2, 61-381 Poznan, Poland e-mail: {jpotoniec,alawrynowicz}@cs.put.poznan.pl |
| Pseudocode | Yes | 1. Refine every pattern from the previous iteration by adding a single restriction for a variable already existing in the pattern. E.g. consider a pattern {?x a :Train.}, its refinements are (1) {?x a :Train, :Cargo Train.}, (2) {?x a :Train, :Passenger Train}, (3) {?x a :Train; :has Engine ?y}. 2. Evaluate patterns (e.g. with some quality measure or using a strategy proposed below) and select only the best ones. 3. Repeat the steps 1-2 as long there are patterns for refinement and maximal number of iterations is not exceeded. |
| Open Source Code | No | The paper does not explicitly state that the source code for their methodology is made publicly available. |
| Open Datasets | Yes | We did preliminary experiments for a few classes from DBpedia ontology, which generated interesting subhierarchies. |
| Dataset Splits | No | The paper does not provide specific training/test/validation dataset splits or cross-validation details. |
| Hardware Specification | No | The paper does not explicitly describe the hardware used for running its experiments. |
| Software Dependencies | No | The paper mentions SPARQL queries but does not list any specific software dependencies with version numbers for reproducibility. |
| Experiment Setup | Yes | In Fr-ONT-Qu s declarative bias, only properties specific for the root class are used (e.g. for the DBpedia class Populated Place, the property province is used, whereas highest Place is not). Data can sometimes be erroneous, so to denoise generated refinements, a minimal coverage threshold is applied. At the same time, we require that at least a given number of patterns is selected (usually 2 or 3), to avoid patterns that do not divide the taxonomy. |