SciProvMiner: captura e consulta de proveniência utilizando recursos Web semânticos para ampliação do conhecimento gerado e otimização do processo de coleta

Abstract

To provide historical scientific information to deal with knowledge loss about scientific experiment has been the focus of several researches. However, the computational support for scientific experiment on a large scale is still incipient and is considered one of the challenges set by the Brazilian Computer Society for 2006 to 2016 period. This work aims to contribute in this area, presenting the SciProvMiner architecture, which main objective is to collect prospective and retrospective provenance of scientific experiments, using ontologies and inference engines to provide useful information in order to increase the knowledge of scientists about a given experiment. We can highlight as specific contributions of SciProvMiner: - Development of a model that encompass prospective and retrospective provenance as an extension of the Open Provenance Model (OPM), which originally onlydeals with retrospective provenance. - Specification and implementation of a provenance collector that uses Web services technology to capture both types of provenance (prospective and retrospective) according to the above model; - Development of an ontology,named Extended OPMO-e, that extends the Open Provenance Model Ontology (OPMO) in order to model prospective provenance beyond the retrospective provenance already covered in OPMO and where are implemented inference and completeness rules defined in OPM documentation. These rules increase the knowledge of scientists on the experiment inferring information that were not explicitly provided by the user and making it possible to optimize the provenance capture mechanism and the consequent decrease on scientist work in order to instrument the workflow. - A relational database specification, where captured provenance information are stored. These information can be used to formulate queries about the provenance explicitly captured, besides provide data to other functionalities of SciProvMiner.