IPB University Logo

SCIENTIFIC REPOSITORY

IPB University Scientific Repository collects, disseminates, and provides persistent and reliable access to the research and scholarship of faculty, staff, and students at IPB University

AI Repository
 
Building and Categories


      View Item 
      •   IPB Repository
      • Final Assignments
      • Undergraduate Final Assignments
      • UF - School of Data Science, Mathematic and Informatics
      • UF - Statistics and Data Sciences
      • View Item
      •   IPB Repository
      • Final Assignments
      • Undergraduate Final Assignments
      • UF - School of Data Science, Mathematic and Informatics
      • UF - Statistics and Data Sciences
      • View Item
      JavaScript is disabled for your browser. Some features of this site may not work without it.

      Perbandingan Metode Random Survival Forest dan XGBoost-AFT dalam Memprediksi Ketahanan Lagu pada Spotify Weekly Chart Indonesia

      Thumbnail
      View/Open
      Cover (461.0Kb)
      Fulltext (1.233Mb)
      Lampiran (498.4Kb)
      Date
      2026
      Jenis/Type
      Skripsi
      Subtype
      Undergraduate Theses
      Author
      Erira, Salsa Rifda
      Indahwati
      Oktarina, Sachnaz Desta
      Metadata
      Show full item record
      Abstract
      Tangga musik digital merupakan indikator penting dalam mengukur kesuksesan lagu di era streaming. Penelitian ini bertujuan untuk membandingkan algoritma Random Survival Forest (RSF) dan Extreme Gradient Boosting–Accelerated Failure Time (XGBoost-AFT) dalam memprediksi ketahanan lagu pada Spotify Weekly Chart Indonesia tahun 2020-2024 dengan 1193 amatan. Kejadian didefenisikan sebagai lagu keluar dari tangga musik, sementara lagu yang masih bertahan hingga akhir penelitian diperlakukan sebagai data tersensor. Validasi model dilakukan menggunakan repeated stratified 5-fold cross validation dengan metrik concordance index (C-index), C-index Inverse Probability of Censoring Weighting (C-Index IPCW), dan mean time-dependent Area Under Curve (mean time-dependent AUC). Hasil penelitian menunjukkan bahwa model XGBoost-AFT memiliki performa yang sedikit lebih baik dibandingkan RSF dengan nilai C-index, C-index IPCW, dan mean time-dependent AUC masing-masing sebesar 0,6440; 0,6435; dan 0,6932. Analisis peubah penting menggunakan feature importance dan SHAP menunjukkan bahwa genre dan bahasa lagu merupakan peubah paling penting, diikuti oleh internal competition, danceability, keberadaan music video (MV), speechiness, dan valence. Selain itu, karakteristik yang diasosiasikan dengan ketahanan lebih panjang antara lain nilai internal competition yang rendah, acousticness tinggi, danceability moderat, adanya MV, serta speechiness dan valence yang relatif rendah, dengan durasi lagu sekitar tiga menit.
       
      Digital music charts are an important indicator of song success in the streaming era. This study compares Random Survival Forest (RSF) and Extreme Gradient Boosting–Accelerated Failure Time (XGBoost-AFT) algorithms in predicting song longevity on the Spotify Weekly Chart Indonesia from 2020 to 2024, using 1,193 observations. The event is defined as a song exiting the chart, while songs that remain until the end of the observation period are treated as censored data. Model validation uses repeated stratified 5-fold cross-validation with evaluation metrics including the concordance index (C-index), C-index Inverse Probability of Censoring Weighting (C-index IPCW), and mean time-dependent Area Under the Curve (mean time-dependent AUC). Results show that XGBoost-AFT performs slightly better than RSF, with C-index, C-index IPCW, and mean time-dependent AUC values of 0.6440, 0.6435, and 0.6932, respectively. Feature importance and SHAP analysis show that genre and song language are the most influential variables, followed by internal competition, danceability, music video (MV) availability, speechiness, and valence. Songs with lower internal competition, higher acousticness, moderate danceability, MV availability, and lower speechiness and valence, with durations around three minutes, tend to have longer chart longevity.
       
      URI
      http://repository.ipb.ac.id/handle/123456789/176630
      Collections
      • UF - Statistics and Data Sciences [166]

      Copyright © 2020 Library of IPB University
      All rights reserved
      Contact Us | Send Feedback
      Indonesia DSpace Group 
      IPB University Scientific Repository
      UIN Syarif Hidayatullah Institutional Repository
      Universitas Jember Digital Repository
        

       

      Browse

      All of IPB RepositoryCollectionsBy Issue DateAuthorsTitlesSubjectsThis CollectionBy Issue DateAuthorsTitlesSubjects

      My Account

      Login

      Application

      google store

      Copyright © 2020 Library of IPB University
      All rights reserved
      Contact Us | Send Feedback
      Indonesia DSpace Group 
      IPB University Scientific Repository
      UIN Syarif Hidayatullah Institutional Repository
      Universitas Jember Digital Repository