OPTIMA: Framework Selecting Optimal Virtual Model to Query Large Heterogeneous Data

OPTIMA is a framework that enables querying the original data on-the-fly without any materialization. It implements two different virtual data models, GRAPH and TABULAR, to join and aggregate data. OPTIMA leverages ontology-based data access and calls the deep learning method to predict the optimal virtual data model using the features extracted from SPARQL queries. Extensive experiments show a reduction in query execution time of over 40% for the TABULAR model selection, and over 30% for the GRAPH model selection.