Open access Journal SINTA 3
Analisis Sentimen Layanan Pelanggan Provider Internet dengan Algoritma Support Vector Machine dan Naïve Bayes
SemanTIK: Teknik Informasi · Vol. 11 No. 1 (2025)
- ISSN
- 2502-8928, 2460-1446
- Published
- 2025-06-19
- Publisher
- Informatics Engineering Department of Halu Oleo University
- Language
- eng
Abstract
Meningkatnya keluhan dan pujian pelanggan terhadap layanan internet menunjukkan pentingnya memahami opini publik secara menyeluruh. Jika hal ini tidak dimanfaatkan dengan baik, perusahaan provider internet dapat kehilangan peluang untuk memperbaiki layanan secara tepat sasaran, yang pada akhirnya menurunkan kepuasan pelanggan. Penelitian ini bertujuan menganalisis sentimen pelanggan terhadap perusahaan penyedia layanan internet Biznet menggunakan Text Mining dengan algoritma Naïve Bayes dan Support Vector Machine (SVM). Data diambil dari X.com (Twitter) menggunakan Tweepy, lalu diproses melalui pembersihan, normalisasi, tokenisasi, pelabelan sentimen dengan VADER, serta ekstraksi fitur menggunakan TF-IDF (Term Frequency–Inverse Document Frequency) untuk mengubah teks menjadi data numerik yang relevan. Hasil menunjukkan bahwa SVM dengan kernel Sigmoid memberikan akurasi tertinggi (94,29%), diikuti oleh SVM Linear (93,92%) dan Naïve Bayes Bernoulli (88,21%). Berdasarkan akurasi, precision, recall, dan F1-score, SVM dengan kernel Sigmoid menjadi metode terbaik untuk analisis sentimen layanan perusahaan penyedia internet Biznet. Analisis sentimen ini diharapkan dapat membantu perusahaan memahami opini pelanggan secara lebih mendalam, sehingga perusahaan provider internet dapat menyusun strategi peningkatan layanan yang tepat sesuai kebutuhan pengguna. The increasing number of customer complaints and compliments regarding internet services highlights the importance of thoroughly understanding public opinion. If not properly leveraged, Biznet company may miss opportunities to improve their services effectively, ultimately reducing customer satisfaction. This study aims to analyze customer sentiment toward Biznet using Text Mining techniques with the Naïve Bayes and Support Vector Machine (SVM) algorithms. Data was collected from X.com (Twitter) using Tweepy, then processed through cleaning, normalization, tokenization, sentiment labeling using VADER, and feature extraction using TF-IDF (Term Frequency–Inverse Document Frequency) to convert text into relevant numerical data. The results show that the SVM with a Sigmoid kernel achieved the highest accuracy (94.29%), followed by the Linear SVM (93.92%) and Bernoulli Naïve Bayes (88.21%). Based on accuracy, precision, recall, and F1-score, the SVM with a Sigmoid kernel is the best method for sentiment analysis of internet provider company’s services. This sentiment analysis is expected to help the company gain deeper insights into customer opinions, enabling the development of targeted strategies for service improvement that align with user needs.
Keywords
Collaboration network of the authors
13 authors · 24 co-authorship links · 1 across different organisations
Hover or tap a name to highlight its co-authors. Double-click to open the author's page.
- Universitas Muhammadiyah Sidoarjo
- Universitas Jambi
- Affiliation unknown
- Different organisations
- Same organisation
- Affiliation unknown
- Thicker line = more joint articles · bigger circle = more articles
Show as table
| Co-author | Organisation | Joint articles |
|---|---|---|
| Novia Ariyanti | Universitas Muhammadiyah Sidoarjo | 2 |
| Gita Wardani | Universitas Muhammadiyah Sidoarjo | 2 |
| Naila Farah Diba | Universitas Muhammadiyah Sidoarjo | 2 |
| Cindy Cahyaning Astuti | Universitas Muhammadiyah Sidoarjo | 1 |
| Herlinda Maya Kumala Sari | Universitas Muhammadiyah Sidoarjo | 1 |
| Uce Indahyanti | Universitas Muhammadiyah Sidoarjo | 1 |
| Vevy Liansari | Universitas Muhammadiyah Sidoarjo | 1 |
| Alfan Indra Kusuma | Universitas Muhammadiyah Sidoarjo | 1 |
| Sumarno Sumarno | Universitas Jambi | 1 |
Affiliations come from OpenAlex; journal OAI-PMH metadata carries none. Authors without an OpenAlex ID are matched by name, which can merge or split people.