Modeling Indonesian Thesaurus on Simple Knowledge Organization System

This paper details the transformation of the Indonesian thesaurus Tesamoko into a SKOS ontology. The process involves scanning the Tesamoko book, performing OCR, and extracting lemmas, absorption words, word classes, and language varieties mapped into SKOS. We introduce new ontology vocabularies to enhance the SKOS Indonesian thesaurus model. The evaluation of SKOS focuses on its ontology aspects and its implementation in information retrieval (IR). Results indicate that the SKOS thesaurus model needs improvement, as it only meets the conciseness criterion. Four experimental setups were used to assess the thesaurus's effectiveness in IR, involving the thesaurus for query and corpus expansion. The F1 measure shows that thesaurus utilization does not enhance IR performance; without thesaurus expansion, the F1 score is 0.218, while with thesaurus expansion, it drops below 0.15. However, corpus expansion performs slightly better than query or combined query and corpus expansion.