Aspect-Based Sentiment Classification of iPhone 15 YouTube Reviews Using VADER-Augmented LSTM
-
Hasna Rafida Alya
*
Telkom University, Bandung,
Indonesia
- Yuliant Sibaroni Telkom University, Bandung, Indonesia (*) Corresponding Author
Abstract
This research investigates the effectiveness of the Long Short-Term Memory (LSTM) model in performing aspect-based sentiment classification on English-language reviews of the iPhone 15 sourced from the YouTube platform. The study focuses on five key product aspects frequently mentioned by users: charger port, camera, screen, design, and battery. To evaluate the model’s performance, two distinct labeling strategies were employed. The first involved manual annotation, where human annotators identified both the relevant aspects and the associated sentiment in each review. The second strategy integrated additional sentiment cues derived from a lexicon-based method, Valence Aware Dictionary and sEntiment Reasoner (VADER). In this approach, the polarity output from VADER was prepended to each review to enrich the input with emotional context. The experimental results demonstrate that supplementing review texts with sentiment polarity information from VADER contributes to a modest but measurable improvement in sentiment classification accuracy. Specifically, using the micro-average accuracy metric, defined as the ratio of correct predictions to the total number of test instances, the model's performance improved from 67% under the manual only annotation to 68% with VADER enhanced input. Additionally, aspect classification remained consistently strong, showing a slight improvement from 90% to 91% after incorporating VADER. Furthermore, based on macro-average accuracy an evaluation metric that calculates the mean performance across all classes regardless of class distribution, accuracy improvements were observed in several aspects, particularly the camera, screen, and design. However, a minor decline in performance was noted for the battery and charger port aspects. These results suggest that enriching review data with sentiment polarity information derived from lexicon-based tools like VADER can enhance the model’s ability to comprehend emotional nuance, leading to more accurate identification of user sentiments within aspect-specific reviews.
Downloads
References
N. Niessen, “Shot on iPhone: Apple’s World Picture,” Advertising & Society Quarterly, vol. 22, no. 2, 2021, doi : https://dx.doi.org/10.1353/asr.2021.0023.
Counterpoint, “ iPhone 15 World’s Best-selling Smartphone in Q3 2024,” Counterpoint Research, 2024. [Online]. Available: https://www.counterpointresearch.com/insight/top-10-bestselling-smartphone-q3-2024/.
I. N. Sandri and Y. Sibaroni, “Perbandingan Model LSTM GloVe dengan LSTM Word2Vec dalam Analisis Sentimen Layanan Dompet Digital,” Telkom University Repository, Sep. 2021.
W. Astriningsih and D. Hatta Fudholi, “Multi Aspect Sentiment Analysis in Hotel Review Using Deep Learning,” Jurnal Teknik Informatika Dan Sistem Informasi (JATISI), vol. 10, no. 3, 2023, doi : https://doi.org/10.35957/jatisi.v10i3.5321.
D. Wahyuni, N. Fadhillah, and W. W. Ariestya, “Long Short-Term Memory dan Lexicon Based Untuk Analisis Sentimen Ulasan Aplikasi TikTok,” Jurnal Ilmiah Komputasi, vol. 23, no. 2, Jun. 2024, doi: 10.32409/jikstik.23.2.3579.
J. Setiawan, V. Gousander, and I. Prasetiawan, “Unmasking the Sentiments of Labuan Bajo: An Instagram-based Analysis for Tourism Insights through VADER Sentiment Analysis,” G-Tech: Jurnal Teknologi Terapan, vol. 7, no. 3, pp. 967–976, Jul. 2023, doi: 10.33379/gtech.v7i3.2615.
M. Gultom, J. Marikros, and W. Rusli, “Penerapan Vader Sentiment untuk Mendeteksi Sentimen Bahasa Inggris berbasis Website,” Seminar Nasional Penelitian (SEMNAS CORISINDO), pp. 13-18, 2024.
G. Arum Prabowo, B. Rahmat, H. Endah Wahanani, U. Pembangunan Nasional Veteran Jawa Timur, and J. Raya Rungkut Madya Gunung Anyar Surabaya, “Aspect-Based Sentiment Analysis iPhone 14 Pro Menggunakan Algoritma XGBoost,” Jurnal Mahasiswa Teknik Informatika (JATI), vol. 7, no. 9, Des. 2023, doi : https://doi.org/10.36040/jati.v7i6.7831.
R. Said and A. Faraby, “Perbandingan Algoritma Machine Learning untuk Analisis Sentimen Berbasis Aspek pada Review Female Daily,” eProceedings of Engineering, vol. 10, no. 3, pp. 3591–3600, 2023.
M. H. Al-Areef and K. Saputra, “Analisis Sentimen Pengguna Twitter Mengenai Calon Presiden Indonesia Tahun 2024 Menggunakan Algoritma LSTM,” Jurnal Sains Manajemen Informatika dan Komputer (SAINTIKOM), vol. 22, no. 2, pp. 270–279, 2023, doi : https://doi.org/10.53513/jis.v22i2.8680.
A. D. Widiantoro, Mustafid and R. Sanjaya, “Pengantar NLP dan Topik Model LDA Sampul Dalam,” Asosiasi Doktor Sistem Informasi Indonesia, Nov. 2024.
I. H. Kusuma and N. Cahyono, “Analisis Sentimen Masyarakat Terhadap Penggunaan E-Commerce Menggunakan Algoritma K-Nearest Neighbor,” Jurnal Pengembangan IT, vol. 8, no. 3, 2023, doi 10.30591/jpit.v8i3.5734.
Y. Amalia, N. Jannah, and R. B. Prasetyo, “Analisis Sentimen dan Emosi Publik pada Awal Pandemi COVID-19 Berdasarkan Data Twitter dengan Pendekatan Berbasis Leksikon,” Seminar Nasional Official Statistics. Vol. 2022. No. 1, pp. 597-608, Nov. 2022, doi : https://doi.org/10.34123/semnasoffstat.v2022i1.1483.
F. Fazrin, O. N. Pratiwi, and R. Andreswari, “Perbandingan Algoritma K-Nearest Neighbor dan Logistic Regression pada Analisis Sentimen terhadap Vaksinasi Covid-19 pada Media Sosial Twitter dengan Pelabelan Vader dan Textblob,” eProceedings of Engineering, vol. 10, no.2 Apr. 2023, Telkom University.
R. Refianti, A. B. Mutiara, and R. A. Putra, “A Lexicon-Based Long Short-Term Memory (LSTM) Model for Sentiment Analysis to Classify Halodoc Application Reviews on Google Playstore,” Journal of Applied Data Sciences, vol. 5, no. 1, pp. 146–157, Jan. 2024, doi: 10.47738/jads.v5i1.160.
H. Bichri, A. Chergui, and M. Hain, “Investigating the Impact of Train / Test Split Ratio on the Performance of Pre-Trained Models with Custom Datasets,” International Journal of Advanced Computer Science and Applications (IJACSA), vol. 15, no. 2, 2024, doi : 10.14569/IJACSA.2024.0150235.
D. M. Gunarto, S. Sa, and D. Q. Utama, “Predicting Cryptocurrency Price Using RNN and LSTM Method,” Jurnal Sistem Informasi dan Komputer, vol. 12, pp. 1–8, 2023, doi: 10.32736/sisfokom.v10i3.1554.
K. Sofi, A. S. Sunge, S. R. Riady, and A. Z. Kamalia, “Perbandingan Algoritma Linear Regression, LSTM, dan GRU dalam Memprediksi Harga Saham dengan Model Time Series,” SEMINASTIKA, vol. 3, no. 1, pp. 39–46, Nov. 2021, doi: 10.47002/seminastika.v3i1.275.
D. R. Alghifari, M. Edi, and L. Firmansyah, “Implementasi Bidirectional LSTM untuk Analisis Sentimen Terhadap Layanan Grab Indonesia,” Jurnal Manajemen Informatika (JAMIKA), vol. 12, no. 2, pp. 89–99, Sep. 2022, doi: 10.34010/jamika.v12i2.7764.
N. A. Dirfas and V. R. S. Nastiti, “Perbandingan Kinerja Pre-Trained Word Embedding Terhadap Performa Klasifikasi Sentimen Ulasan Produk Tokopedia Dengan Long Short-Term Memory(LSTM),” Building of Informatics, Technology and Science (BITS), vol. 6, no. 2, Sep. 2024, doi: 10.47065/bits.v6i2.5634.
W. Aljedaani, F. Rustam, S. Ludi, A. Ouni, and M. W. Mkaouer, “Learning Sentiment Analysis for Accessibility User Reviews,” 2021 36th IEEE/ACM International Conference on Automated Software Engineering Workshops (ASEW), 2021, pp. 239–246. doi: 10.1109/ASEW52652.2021.00053.
Bila bermanfaat silahkan share artikel ini
Berikan Komentar Anda terhadap artikel Aspect-Based Sentiment Classification of iPhone 15 YouTube Reviews Using VADER-Augmented LSTM
Pages: 685-695
Copyright (c) 2025 Hasna Rafida Alya, Yuliant Sibaroni
This work is licensed under a Creative Commons Attribution 4.0 International License.
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under Creative Commons Attribution 4.0 International License that allows others to share the work with an acknowledgment of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgment of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (Refer to The Effect of Open Access).
