Explainability of text Classification through ontology-driven analysis in Serious Games

Anar Mammadli, Elviz Ismayilov, Cecilia Zanni-Merk

Abstract: Several approaches have been proposed to study text Classification in serious games. Unfortunately, those works do not investigate the impact of reasoning on text-based Classification. This paper proposes a novel approach to give ontology-driven reasoning on predicted results by machine learning. Thus, we construct text embedding on the serious game data. Next, we use machine learning for prediction and DBpedia ontology for reasoning to analyze the enhancement in prediction results. We conduct a set of benchmarks using several well-known machine learning algorithms, leading to an accuracy of 97%.