Check out our talk on "Semantic Annotation of Predictive Modeling Experiments" by Ilin Tolovski at the 23rd International Conference on Discovery Science (free online participation). The talk is scheduled in the Data and Knowledge Representation Session on 20th of October at 17:00.
Executing various predictive modeling experiments is a wide-spread practice, both in industry and academia. This results in a production of computational models and experiment artefacts that strain both the computational and financial resources of an institution. Here arises the challenge of proper representation and storage of experimental setups, and results. We introduce OntoExp, an OntoDM module which gives a granular representation of a predictive modeling experiment that enables annotation of the experiment's provenance, algorithm implementations, parameter settings and output metrics. This module is incorporated in SemanticHub, an online system that allows users to either execute their own experiment or browse the repository of completed experimental workflows.
We showcase the capabilities of the system with executing multi-target regression experiment on water quality prediction in the CLUS data mining framework. The system and created repositories are evaluated based on the FAIR data stewardship guidelines. The evaluation shows that OntoExp and SemanticHub provide the infrastructure needed for semantic annotation, execution, storage, and querying of the experiments. As a result, the created experiment repositories are adequately represented according to the FAIR principles for data stewardship.
The full paper is available on this link and will be published in the Proceedings of the 23rd International Conference on Discovery Science .