Sistem Ektraksi Informasi Lirik Lagu Daerah Menggunakan Term Frequency – Inverse Document Frequency Dan Cosine Similarity

Traditional regional songs are one of Indonesia’s cultural heritages that contain historical values and local wisdom. However, many regional song lyrics are still difficult to access and widely understand. This study aims to develop an information extraction system for regional song lyrics using the Term Frequency–Inverse Document Frequency (TF-IDF) and Cosine Similarity methods to support keyword-based lyric searching. The system was developed through text preprocessing stages, term weighting using TF-IDF, and similarity measurement between user queries and lyric documents using Cosine Similarity. In addition, the system is equipped with linguistic analysis using POS Tagging and rule-based extraction to identify elements such as nouns, verbs, subjects, objects, negation, and emotions. The test results show that the system achieved a Top-1 Accuracy of 100% on 10 test queries. The linguistic extraction evaluation produced a precision of 0.74, recall of 0.76, and an F1-score of 0.75. Therefore, this system is able to effectively extract information from regional song lyrics as a tool for cultural preservation and digital education.