ANALISIS SENTIMEN BERBASIS ASPEK PADA ULASAN APLIKASI KAI ACCESS MENGGUNAKAN METODE SUPPORT VECTOR MACHINE
Article Sidebar
Main Article Content
Abstract
PT Kereta Api Indonesia melakukan inovasi dengan meluncurkan aplikasi yang diberi nama KAI Access. Aplikasi KAI Access memiliki fitur pemesanan tiket, ticket rescheduling, pembatalan tiket hingga e-boarding pass. Tujuan dari penelitian ini adalah untuk mengetahui sentimen dari sebuah produk mobile. Opini terkait Aplikasi KAI Access dapat digunakan PT Kereta Api Indonesia sebagai parameter kunci untuk mengetahui tingkat kepuasan publik sekaligus bahan evaluasi bagi PT Kereta Api Indonesia. Berdasarkan hasil pengujian yang telah dilakukan pada ulasan pengguna aplikasi KAI Access dengan total 8.078 ulasan, lebih banyak pengguna memberikan opini positif dalam aspek satisfaction dan opini negatif pada aspek learnability, efficiency, dan errors. Digunakan model CRISP-DM (Cross Industry Standard Process for Data Mining) dan algoritma Support Vector Machine untuk melakukan klasifikasi. Hasil klasifikasi terbaik diperoleh nilai accuracy, precision, recall, dan F-measure yang dihasilkan dari tiap aspek yaitu untuk Learnability 94.73%, 100.00%, 89.50%, dan 94.64%, Efficiency 94.38%, 72.00%, 100.00%, dan 94.46%, Errors 85.13%, 97.11%, 72.41%, dan 82.96%, Satisfaction 87.26%, 98.46%, 73.78%, dan 84.20%.
PT Kereta Api Indonesia innovates by launching an application called KAI Access. The KAI Access application has features for ticket ordering, ticket rescheduling, ticket cancellation and e-boarding pass. The purpose of this study is to determine the sentiment of a mobile. Opinion regarding the KAI Access Application can be used by PT Kereta Api Indonesia as a key parameter to determine the level of public satisfaction as well as evaluation material for PT Kereta Api Indonesia. Based on the results of tests conducted on user reviews of the KAI Access application with a total of 8,078 reviews, more users give positive opinions on the satisfaction and negative opinions on the learnability, efficiency and errors. Model CRISP-DM (Cross Industry Standard Process for Data Mining) and Support Vector Machine algorithm are used to perform classification. The best classification results obtained accuracy, precision, recall, and F-measure resulting from each aspect, namely for Learnability 94.73%, 100.00%, 89.50%, and 94.64%, Efficiency 94.38%, 72.00%, 100.00%, and 94.46%, Errors 85.13%, 97.11%, 72.41%, and 82.96%, Satisfaction 87.26%, 98.46%, 73.78%, and 84.20%.
Downloads
Article Details
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
JUKANTI Journal License
JUKANTI is committed to promoting open access and the free distribution of knowledge. We implement the following license model to ensure fair and ethical use of the materials published.
Creative Commons Attribution 4.0 International License (CC BY 4.0)
All articles published by JUKANTI are licensed under the Creative Commons Attribution 4.0 International License. This license allows users to:
- Copy and Distribute: Users are free to copy, distribute, and display the original work, provided they give appropriate credit to the authors and the source.
- Adapt: Users can modify, change, and build upon the original work, provided they give appropriate credit and indicate if changes were made.
- Commercial Use: Users can use the work for commercial purposes, provided they give appropriate credit.
Author Obligations
Authors publishing their articles with JUKANTI agree to:
- Guarantee that the work is original and free from copyright infringement.
- Grant permission to JUKANTI to publish the work under the CC BY 4.0 license.
- Retain the original copyright of their work, with the publication license granted to JUKANTI.
Compliance with DOAJ
JUKANTI is committed to complying with the guidelines and standards set by the Directory of Open Access Journals (DOAJ). We strive to ensure integrity, transparency, and high quality in all our publications.
For further questions or clarifications regarding this license, please contact Jukanti Editor at jukanti.ejournalcbn@gmail.com.
