AloLex: A Linked Open Data Lexicon for Abusive Language in Serbian

AloLex is a Linked Open Data lexicon for abusive language in Serbian represented in RDF/Turtle and modeled using the OntoLex-Lemon framework. The resource was constructed through the integration and harmonization of twelve heterogeneous lexical resources containing abusive language expressions, hate speech terms, profanity, threats, slang expressions, and related lexical material. The resource contains 11,489 lexical entries, including 7,212 single-token words and 4,277 multi-word units. AloLex preserves provenance information for lexical entries and includes additional linguistic and semantic annotations such as grammatical information, abusive language categories, hate speech categories, meanings, synonyms, and usage examples. This Zenodo record contains the AloLex RDF/Turtle resource together with supplementary materials used in its evaluation, including annotated abusive-language examples and evaluation annotations. Due to platform restrictions, original Twitter-based examples are not redistributed in the public release.