Learning from Ontology Streams with Semantic Concept Drift

TL;DR

This paper introduces a method for handling concept drift in ontology streams by leveraging semantic inference and embeddings, improving prediction accuracy in dynamic data environments.

Contribution

It presents a novel approach that combines semantic inference and embeddings to address concept drift in ontology stream learning.

Findings

Accurate predictions on Dublin and Beijing data streams.

Semantic methods outperform traditional models in concept drift scenarios.

Effective handling of unexpected data distribution changes.

Abstract

Data stream learning has been largely studied for extracting knowledge structures from continuous and rapid data records. In the semantic Web, data is interpreted in ontologies and its ordered sequence is represented as an ontology stream. Our work exploits the semantics of such streams to tackle the problem of concept drift i.e., unexpected changes in data distribution, causing most of models to be less accurate as time passes. To this end we revisited (i) semantic inference in the context of supervised stream learning, and (ii) models with semantic embeddings. The experiments show accurate prediction with data from Dublin and Beijing.