The IMPLEMENTASI  BERT UNTUK ANALISIS SENTIMEN PENGGUNA GAME ONLINE E-FOOTBALL PADA ULASAN DI PLAY STORE  BERBASIS MACHINE LEARNING

Muhammad Aryaraka Pradana; Fido Rizki; Cindi Wulandari

[1] R. Adolph, “済無No Title No Title No Title,” pp. 1–23, 2016.
[2] R. D. Wahyuni and A. N. Utomo, “Penggunaan Metode Lexicon Untuk Analisis Sentimen pada Ulasan Aplikasi KAI Access di Google Play Store,” J. Rekayasa Inf., vol. 11, no. 2, pp. 134–145, 2022.
[3] R. Mas, R. W. Panca, K. Atmaja1, and W. Yustanti2, “Analisis Sentimen Customer Review Aplikasi Ruang Guru dengan Metode BERT (Bidirectional Encoder Representations from Transformers),” Jeisbi, vol. 02, no. 3, pp. 55–62, 2021.
[4] M. K. Insan, U. Hayati, and O. Nurdiawan, “Analisis Sentimen Aplikasi Brimo Pada Ulasan Pengguna Di,” J. Mhs. Tek. Inform., vol. 7, no. 1, pp. 478–483, 2023.
[5] Fajri91, “InSet (Indonesia Sentiment Lexicon),” Github.Com, 2018, [Online]. Available: https://github.com/fajri91/InSet

Abstract viewed - 5 times
PDF downloaded - 5 times

Affiliations

Muhammad Aryaraka Pradana
Universitas Bina Insan

Fido Rizki
Universitas Bina Insan

Cindi Wulandari
Universitas Bina Insan

Content Top

The IMPLEMENTASI BERT UNTUK ANALISIS SENTIMEN PENGGUNA GAME ONLINE E-FOOTBALL PADA ULASAN DI PLAY STORE BERBASIS MACHINE LEARNING

Studi Sentimen Pengguna terhadap E-Football di Google Play Store Menggunakan Model BERT

Muhammad Aryaraka Pradana ,
Fido Rizki ,
Cindi Wulandari

Proceedings Economic, Social Science, Computer, Agriculture and Fisheries (ESCAF) 4th 2025

Submitted: Jan 30, 2025

Abstract

Abstrak

Analisis sentimen pada ulasan pengguna dapat memberikan wawasan penting terkait persepsi dan kepuasan pengguna terhadap suatu aplikasi, termasuk game online E-Football. Penelitian ini bertujuan mengimplementasikan model BERT untuk analisis sentimen ulasan pengguna game online E-Football di Google Play Store. Dataset terdiri dari 25.000 ulasan dengan tiga klasifikasi sentimen: Positif, Netral, dan Negatif. Ulasan diproses melalui Cleaning, tokenization, case folding, stopwords, normalization, dan stemming. Model BERT dilatih dan dievaluasi menggunakan metrik akurasi, presisi, recall, dan f1-score. Hasil menunjukkan akurasi sebesar 90%. Untuk kelas negatif, presisi 95% dan recall 97%, kelas positif mencapai presisi 87% dan recall 85%. Namun, kelas netral memiliki performa lebih rendah dengan presisi 58% dan recall 53%.Kesimpulan menunjukkan bahwa model BERT efektif untuk analisis sentimen dalam bahasa Indonesia, dengan peluang pengembangan lebih lanjut untuk meningkatkan performa analisis sentimen pada kelas netral dan aplikasi pada domain lainnya.

Kata kunci : Analisis Sentimen, BERT, Google Play Store, E-Football

Abstract

Sentiment analysis on user reviews can provide important insights into user perception and satisfaction with an application, including the online game E-Football. This study aims to implement the BERT model for sentiment analysis of user reviews of the E-Football online game on Google Play Store. The dataset consists of 25,000 reviews categorized into three sentiment classes: Positive, Neutral, and Negative. The reviews underwent preprocessing, including data cleaning, tokenization, case folding, stopword removal, normalization, and stemming. The BERT model was trained and evaluated using accuracy, precision, recall, and f1-score metrics. The results show an accuracy of 90%. For the negative class, the precision is 95% with a recall of 97%; the positive class achieved 87% precision and 85% recall. However, the neutral class showed lower performance, with 58% precision and 53% recall. In conclusion, the BERT model is effective for sentiment analysis in the Indonesian language. There are further opportunities to enhance sentiment analysis performance for the neutral class and to apply the model in other domains.

Keywords: Sentiment Analysis, BERT, Google Play Store, E-Football