EUVIMEDEUVIMED entdecken
Crossref

KLASIFIKASI SENTIMEN MASYARAKAT TERHADAP KASUS TUNTUTAN 17+8 MENGGUNAKAN NAÏVE BAYES CLASSIFIER

Arif Kurniawan, Muhammad Fikry, Novi Yanti, Surya Agustian

EXPERT: Jurnal Manajemen Sistem Informasi dan Teknologi · 30.06.2026 · 16 (1) · 81

DOI: 10.36448/expert.v16i1.4947 ↗

Kurz gefasst

Fachartikel in der Zeitschrift EXPERT: Jurnal Manajemen Sistem Informasi dan Teknologi, online erschienen am 30.06.2026. Verfasst von Arif Kurniawan und 3 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

– The rapid growth of social media has encouraged the emergence of diverse public opinions on various public issues, including the 17+8 Demands case, which has attracted widespread attention in Indonesia. Sentiment analysis serves as an important approach for systematically identifying public opinion trends. However, social media text data are generally unstructured and contain various forms of noise, requiring appropriate preprocessing techniques before classification can be performed. This study aims to analyze public sentiment toward the 17+8 Demands case using the Naïve Bayes Classifier method and to evaluate the impact of preprocessing stages on model performance through an ablation study approach. The research data consisted of TikTok comments processed through several preprocessing stages, including case folding, normalization, stopword removal, and stemming. Subsequently, text features were extracted using the TF-IDF weighting method and classified using the Multinomial Naïve Bayes algorithm. The results indicate that each preprocessing combination produces different classification performances. On the additional dataset consisting of 8,370 comments, the best performance was achieved using the combination of case folding and normalization, resulting in an accuracy of 89.87% and an F1-score of 91.89%. Meanwhile, on the primary dataset consisting of 1,525 comments, the best performance was obtained using the combination of normalization and stopword removal, achieving an accuracy of 81.20% and an F1-score of 80.31%. Evaluation using a confusion matrix demonstrated that the model performed well in sentiment classification, particularly for the positive sentiment class. These findings indicate that selecting appropriate preprocessing techniques plays a significant role in improving sentiment classification performance when using the Naïve Bayes Classifier method.

Zitieren

APA

Kurniawan, A., Fikry, M., Yanti, N., & Agustian, S. (2026). KLASIFIKASI SENTIMEN MASYARAKAT TERHADAP KASUS TUNTUTAN 17+8 MENGGUNAKAN NAÏVE BAYES CLASSIFIER. EXPERT: Jurnal Manajemen Sistem Informasi dan Teknologi, 16(1), 81. https://doi.org/10.36448/expert.v16i1.4947

Vancouver

Kurniawan A, Fikry M, Yanti N, Agustian S. KLASIFIKASI SENTIMEN MASYARAKAT TERHADAP KASUS TUNTUTAN 17+8 MENGGUNAKAN NAÏVE BAYES CLASSIFIER. EXPERT: Jurnal Manajemen Sistem Informasi dan Teknologi. 2026;16(1):81. doi:10.36448/expert.v16i1.4947

BibTeX
@article{Kurniawan2026,
  author = {Kurniawan, Arif and Fikry, Muhammad and Yanti, Novi and Agustian, Surya},
  title = {KLASIFIKASI SENTIMEN MASYARAKAT TERHADAP KASUS TUNTUTAN 17+8 MENGGUNAKAN NAÏVE BAYES CLASSIFIER},
  journal = {EXPERT: Jurnal Manajemen Sistem Informasi dan Teknologi},
  year = {2026},
  volume = {16},
  number = {1},
  pages = {81},
  doi = {10.36448/expert.v16i1.4947}
}
Kostenlos starten →