Linking Emotions: Affective and Lexical Resources for Italian in Linked Open Data

The growing interest in analysing emotions in Italian texts has led to the development of various affective resources, often independently constructed and lacking interoperability. To address this fragmentation, we adopt a Linked Open Data (LOD) approach. This paper presents three main contributions: (1) the release of Sentix 3.0, a revised and enriched polarity lexicon for Italian, together with two derivatives (MAL and WMAL) that address morphological and word frequency variation issues; (2) a new quartile-based methodology to discretize continuous polarity scores; and (3) the linking of Sentix 3. 0 and ELIta, an annotated emotion lexicon with categorical labels, to the LiITA Lemma Bank using standard ontologies (OntoLex-Lemon, MARL) and the newly introduced elita ontology for categorical emotion representation based on Plutchik's Wheel. At the heart of the linking is the word, the central node for aligning different lexical resources.