Machine Learning-Augmented Ontology-Based Data Access for Renewable Energy Data

TL;DR

This paper presents a machine learning-enhanced framework for ontology-based data access in renewable energy, improving data classification and insights from fragmented, imbalanced datasets through dynamic class management.

Contribution

It introduces a novel framework that integrates machine learning with OBDA to handle hierarchical classification and imbalanced data in renewable energy applications.

Findings

Enhanced classification accuracy demonstrated in case studies

Improved data insights from underrepresented classes

Validated effectiveness in real-world renewable energy scenarios

Abstract

Managing the growing data from renewable energy production plants for effective decision-making often involves leveraging Ontology-based Data Access (OBDA), a well-established approach that facilitates querying diverse data through a shared vocabulary, presented in the form of an ontology. Our work addresses one of the common problems in this context, deriving from feeding complex class hierarchies defined by such ontologies from fragmented and imbalanced (w.r.t. class labels) data sources. We introduce an innovative framework that enhances existing OBDA systems. This framework incorporates a dynamic class management approach to address hierarchical classification, leveraging machine learning. The primary objectives are to enhance system performance, extract richer insights from underrepresented data, and automate data classification beyond the typical capabilities of basic deductive reasoning at the ontological level. We experimentally validate our methodology via real-world, industrial case studies from the renewable energy sector, demonstrating the practical applicability and effectiveness of the proposed solution.