Development of a Retrieval System for Indonesian Popular Music Using the Savoy Method

Abstract

This study aims to develop a popular Indonesian music retrieval system using the Savoy method, designed to enhance the effectiveness of finding relevant music documents based on user queries. The implementation process includes stages of indexing, term weighting, and calculating the relevance of music documents, which is done through the calculation of Term Frequency (TF), Inverse Document Frequency (IDF), and TF-IDF for each keyword. Additionally, the system utilizes the Cosine Similarity algorithm to measure the similarity between keywords and music documents in the database. The system is developed as a web-based platform using PHP programming language and MySQL as the database, with the Laravel framework for the user interface. Therefore, this system allows users to easily search for songs based on keywords such as lyrics or song titles and receive relevant song recommendations. The system's testing results show that the developed music retrieval system successfully provides song recommendations with a high level of relevance. Testing with a dataset of 500 popular Indonesian songs shows that the system can identify relevant songs based on user queries with an accuracy of 87%. Furthermore, survey results from 50 respondents indicate that 90% of users are satisfied with the ease of use and relevance of the recommendations provided by the system. This system proves to be effective in improving the efficiency of searching for popular Indonesian music and offering a better user experience.