ANALISIS SENTIMEN PENGGUNA LAYANAN ICONNET DENGAN PENDEKATAN METODE SUPPORT VECTOR MACHINE

Simanjuntak, Junita Ruth Veron and Yosrita, Efy and Luqman, Luqman (2024) ANALISIS SENTIMEN PENGGUNA LAYANAN ICONNET DENGAN PENDEKATAN METODE SUPPORT VECTOR MACHINE. Diploma thesis, ITPLN.

[thumbnail of 202031241_Junita Ruth Veron_Revisi_Skripsi_Junita Ruth Veron Si.pdf] Text
202031241_Junita Ruth Veron_Revisi_Skripsi_Junita Ruth Veron Si.pdf
Restricted to Registered users only

Download (2MB)

Abstract

Seiring dengan meningkatnya kebutuhan kebutuhan manusia di bidang teknologi, informasi dan komunikasi yang semakin meningkat, maka diperlukan juga peningkatan pada layanan internet. Iconnet merupakan penyedia layanan internet yang memiliki kecepatan internet yang tinggi dan stabil. Namun, pengguna mengeluhkan gangguan koneksi internet. Semakin tinggi kualitas pelayanan yang dapat diterima pelanggan, maka semakin tinggi pula harapan pelanggan akan puas terhadap kualitas pelayanan perusahaan. Sehingga penelitian ini bertujuan untuk membuat dataset dan model algoritma support vector machine serta memperoleh evaluasi tingkat akurasi dalam menganalisis sentimen pengguna layanan Iconnet dari platform media sosial X. Metode Support Vector Machine (SVM) digunakan untuk mengklasifikasikan sentimen pengguna menjadi positif, netral, dan negatif. Data ulasan pengguna dikumpulkan melalui proses crawling dari X menggunakan kata kunci terkait Iconnet, dengan rentang waktu antara Maret 2020 hingga Juli 2024. Tahapan preprocessing meliputi pembersihan teks, tokenisasi, penanganan kata slang, penghapusan stopwords, dan stemming. Selanjutnya, data diberi label sentimen menggunakan kamus leksikon dan dilakukan pembobotan kata dengan TF-IDF. Model SVM dengan kernel sigmoid kemudian dilatih dan dievaluasi menggunakan teknik K-Fold Cross Validation dan confusion matrix. Hasil penelitian menunjukkan bahwa model SVM mencapai akurasi 96%, dengan F1-score untuk sentimen negatif 95%, netral 98%, dan positif 97%.

In the face of increasing human needs in the fields of technology, information, and communication, improvements in internet services are also necessary. Iconnet is an internet service provider known for its high-speed and stable internet connections. However, users have reported experiencing internet connection disruptions. The higher the quality of service perceived by customers, the higher their expectations for satisfaction with the company's services. Therefore, this study aims to create a dataset and a Support Vector Machine (SVM) algorithm model, along with evaluating its accuracy in analyzing the sentiments of Iconnet users on the social media platform X. The SVM method is employed to classify user sentiment into positive, neutral, and negative categories. User review data was collected through a crawling process from X using Iconnet-related keywords, spanning from March 2020 to July 2024. Preprocessing stages include text cleaning, tokenization, handling of slang words, stop word removal, and stemming. Subsequently, sentiment labels are assigned to the data using a lexicon dictionary, and word weighting is performed using TF-IDF. The SVM model with a sigmoid kernel is then trained and evaluated using K-Fold Cross Validation and a confusion matrix. The research results show that the SVM model achieves an accuracy of 96%, with F1-scores for negative sentiment at 95%, neutral at 98%, and positive at 97%.

Item Type: Thesis (Diploma)
Uncontrolled Keywords: Analisis Sentimen, Iconnet, Support Vector Machine, Media Sosial, Evaluasi Layanan. Sentiment Analysis, Iconnet, Support Vector Machine, Social Media, Service Evaluation.
Subjects: Skripsi
Bidang Keilmuan > Teknik Informatika
Divisions: Fakultas Telematika Energi > S1 Teknik Informatika
Depositing User: Sudarman
Date Deposited: 30 Sep 2025 07:35
Last Modified: 30 Sep 2025 07:35
URI: https://repository.itpln.ac.id/id/eprint/1580

Actions (login required)

View Item
View Item