ASPECT BASED SENTIMENT ANALYSIS TERHADAP PERTAMINA PADA PLATFORM X MENGGUNAKAN INDOBERTWEET
Isi Artikel Utama
Abstrak
Penelitian ini bertujuan menganalisis opini publik terhadap Pertamina melalui platform X menggunakan pendekatan Aspect-Based Sentiment Analysis (ABSA) dengan model IndoBERTweet. Data cuitan berbahasa Indonesia dikumpulkan periode September–Oktober 2025 menggunakan tweet-harvest v2.6.1 dengan kata kunci "Pertamina lang:id". Penelitian menggunakan kerangka Knowledge Discovery in Database (KDD) yang mencakup pengumpulan data, preprocessing, ekstraksi aspek, klasifikasi sentimen, dan analisis hasil. Ekstraksi aspek menggunakan pendekatan keyword-based menghasilkan lima aspek: Harga BBM, Kualitas BBM, Pelayanan SPBU, Kebijakan Pertamina, dan Umum. Dari 7.486 cuitan mentah diperoleh 6.642 cuitan bersih, kemudian dihasilkan 8.198 data setelah replikasi multi-aspek. Klasifikasi sentimen dilakukan menggunakan model IndoBERTweet pretrained secara otomatis. Hasil menunjukkan sentimen Negatif mendominasi 45,22% (3.707 data), Netral 42,17% (3.457 data), dan Positif 12,61% (1.034 data). Aspek Kebijakan Pertamina (55,4%) dan Kualitas BBM (55,0%) mencatat proporsi negatif tertinggi, sedangkan aspek Umum memiliki distribusi paling berimbang dengan positif tertinggi (28,2%). Temuan ini digunakan sebagai dasar rekomendasi strategis peningkatan layanan dan komunikasi publik Pertamina.
Rincian Artikel
Artikel ini berlisensiCreative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
Copyright (c) STMIK TIMEReferensi
[1] F. Koto, J. H. Lau, and T. Baldwin, “IndoBERTweet: A pretrained language model for Indonesian Twitter,” in Proc. EMNLP 2021, 2021.
[2] DataReportal, “Digital 2023: Indonesia,” 2023. [Online]. Available: https://datareportal.com/reports/digital-2023-indonesia
[3] R. Halim, Y. Nugroho, and P. I. Santosa, “Sentiment analysis of Indonesian tweets using Bidirectional LSTM,” Procedia Comput. Sci., vol. 179, pp. 476–484, 2021.
[4] M. N. Azhar and M. L. Khodra, “Aspect-based sentiment analysis for Indonesian hotel reviews using BERT,” Procedia Comput. Sci., vol. 179, pp. 455–463, 2021.
[5] N. Said and T. Manik, “Analyzing political sentiment in Indonesian elections using ABSA,” Procedia Comput. Sci., vol. 217, pp. 124–132, 2023.
[6] F. Koto, J. H. Lau, and T. Baldwin, “IndoBERTweet,” Proc. EMNLP 2021, 2021.
[7] A. Kurniawan, M. F. Pratama, and S. Wibowo, “Aspect-based sentiment analysis for Indonesian e-commerce using BERT and CNN,” J. Inf. Syst. Informatics, vol. 4, no. 3, pp. 567–579, 2022.
[8] B. Liu, Sentiment Analysis: Mining Opinions, Sentiments, and Emotions, 2nd ed. Cambridge Univ. Press, 2020.
[9] M. Pontiki et al., “SemEval-2016 Task 5: Aspect based sentiment analysis,” Proc. SemEval-2016, pp. 19–30, 2016.
[10] V. Giudice and V. Cutello, “European social sentiment during energy crises: A neural transformer approach,” Inf. Sci., vol. 578, pp. 786–799, 2021.
[11] H. Sutanto and I. Darmawan, “Aspect extraction and sentiment mapping in Indonesian transportation reviews,” JITEKI, vol. 9, no. 2, pp. 224–234, 2023.
[12] R. Sudarsono and N. Fitriani, “Implementasi KDD untuk Analisis Data Media Sosial,” J. Teknol. Inf. Ilmu Komput., vol. 8, no. 3, pp. 411–420, 2021.
[13] M. Handayani and F. Sudibyo, “Tracking social media sentiment toward fuel policy in Indonesia,” Sustainability, vol. 14, no. 17, p. 10891, 2022.
[14] M. Hidayat and S. Yuliana, “Preprocessing and normalization methods for Indonesian social media sentiment analysis,” IJCCS, vol. 16, no. 2, pp. 112–121, 2022.
[15] J. Devlin, M.-W. Chang, K. Lee, and K. Toutanova, “BERT: Pre-training of deep bidirectional transformers,” Proc. NAACL-HLT 2019, pp. 4171–4186, 2019.
[16] M. Ardiyanto, “Aardiiiiy/indobertweet-base-Indonesian-sentiment-analysis,” Hugging Face, 2024. [Online]. Available: https://huggingface.co/Aardiiiiy