Pragmatic Ontology Evolution: Reconciling User Requirements and Application Performance
In this page you can download all data regarding the evaluation of the Pragmatic Ontology Evolution Framework (POE), described in the paper:
Osborne, F., Motta, E.: Pragmatic Ontology Evolution: Reconciling User Requirements and Application Performance. Submitted to ISWC 2018.
POE is a novel approach able to select the set of new concepts for an evolving ontology which ensure that the resulting version of the ontology i) is consistent with the given requirements, ii) is parametrised with respect to a number of dimensions, and iii) supports effectively relevant computational tasks. This approach also supports users in navigating the space of possible solutions by showing how certain choices, such as limiting the number of concepts or privileging trendy concepts rather than historical ones, would reflect on the application performance.
We tested POE on the task of evolving the Product Marked Code taxonomy and used as instances a dataset of Springer Nature publications including 1,218 books in the 2012-2016 period.
The evaluation had two aims. First, we wanted to compare POE versus alternative baselines from the state of the art, such as the TF-IDF method adopted in Text2Onto and SPRAT. Secondly, we intended to investigate whether optimizing for a certain task would also yield good performance on related ones. For instance, optimizing for similarity computation should intuitively yield good performance on all tasks that consider item similarity.
For any question about POE and the evaluation please contact francesco.osborne@open.ac.uk.