PERBANDINGAN NAÏVE BAYES DAN SUPPORT VECTOR MACHINE TERHADAP ULASAN APLIKASI MyICON+

Lestari, Ni Made Dwi and Fitriani, Yessy and Palupiningsih, Pritasari (2024) PERBANDINGAN NAÏVE BAYES DAN SUPPORT VECTOR MACHINE TERHADAP ULASAN APLIKASI MyICON+. Diploma thesis, ITPLN.

[thumbnail of 202031244_Ni Made Dwi Lestari_Revisi_ Skripsi_Ni Made Dwi Lestari.pdf] Text
202031244_Ni Made Dwi Lestari_Revisi_ Skripsi_Ni Made Dwi Lestari.pdf
Restricted to Registered users only

Download (2MB)

Abstract

Perkembangan teknologi informasi dan komunikasi memicu transformasi dalam akses informasi, terutama melalui aplikasi mobile yang disediakan oleh provider Wi-Fi seperti MyICON+. Aplikasi tersebut dikembangkan oleh perushaan PLN ICON Plus yang memberikan layanan pelanggan secara digital. Namun, meningkatnya jumlah ulasan pengguna menimbulkan tantangan dalam analisis sentimen secara manual, yang dapat menghambat peningkatan layanan. Penelitian ini bertujuan membandingkan dua metode klasifikasi Naïve Bayes dan Support Vector Machine (SVM) untuk menganalisis ulasan aplikasi MyICON+. Menggunakan data yang diambil dari Google Play Store melalui web scraping. Proses dimulai dengan pre-processing teks, pelabelan sentimen menggunakan metode Inset Lexicon-based, dan perhitungan TF-IDF untuk pembobotan kata. Model Naïve Bayes dan SVM kemudian dilatih dan diuji dengan rasio data latih dan uji 80:20. Hasil evaluasi menggunakan Confusion Matrix menunjukkan bahwa SVM memiliki akurasi lebih tinggi (81%) dibandingkan Naïve Bayes (73%). Pada penelitian ini menunjukkan bahwa metode SVM lebih baik dalam mengelompokkan ulasan pengguna. Hal ini memberikan wawasan yang lebih baik untuk meningkatkan kualitas layanan MyICON+.

The development of information and communication technology has triggered a transformation in access to information, especially through mobile applications provided by Wi-Fi providers such as MyICON+. The application was developed by PLN ICON Plus, which provides digital customer services. However, the increasing number of user reviews poses challenges in manual sentiment analysis, which can hinder service improvement. This study aims to compare two classification methods Naïve Bayes and Support Vector Machine (SVM) to analyse MyICON+ app reviews. Using data taken from the Google Play Store through web scraping. The process starts with text pre-processing, sentiment labelling using the Inset Lexicon-based method, and TF-IDF calculation for word weighting. Naïve Bayes and SVM models were then trained and tested with a training and test data ratio of 80:20. Evaluation results using Confusion Matrix show that SVM has higher accuracy (81%) than Naïve Bayes (73%). This study shows that the SVM method is better at categorising user reviews, providing better insights to improve the quality of MyICON+ services.

Item Type: Thesis (Diploma)
Uncontrolled Keywords: Ulasan, Naïve Bayes, Support Vektor Machine (SVM), MyICON+ Reviews, Naïve Bayes, Support Vector Machine (SVM), MyICON+
Subjects: Skripsi
Bidang Keilmuan > Teknik Informatika
Divisions: Fakultas Telematika Energi > S1 Teknik Informatika
Depositing User: Sudarman
Date Deposited: 26 Sep 2025 07:18
Last Modified: 26 Sep 2025 07:18
URI: https://repository.itpln.ac.id/id/eprint/1490

Actions (login required)

View Item
View Item