Identifikasi Ekspresi Idiomatik Menggunakan Distributional Semantic Based Approach dan Truth Discovery

Abstract

Idiomatic expressions are phrases that consist of a sequence of two or more words that have a meaning that cannot be predicted from the meaning of the individual words that compose it. Idiomatic expressions exist in almost all languages ??but are difficult to extract because there is no algorithm that can precisely decipher the structure of idiomatic expressions, so most rule-based machine translation systems generally translate idiomatic expressions by translating word for word their constituents, but the translation results do not produce the true meaning of the idiomatic expression. Based on this problem, the author tries to do research on the identification of the use of idiomatic expressions in Indonesian sentences. First, the author conducts the sentence classification process using BERT to find out whether the sentence contains idiomatic expressions or not. Furthermore, idiomatic expressions are identified based on distributional semantic based approach and then validated automatically using the Truth Discovery method. From the research conducted, the identification of idiomatic expressions in Indonesian sentences using Distributional Semantic Based Approach and Truth Discovery obtained an accuracy of 0.82; precision 1.0; recall 0.64 and f1-score 0.78.