ANALISIS SENTIMEN APLIKASI MASKAPAI PENERBANGAN LION AIR MENGGUNAKAN METODE SVM DAN NAÏVE BAYES (STUDI KASUS:ULASAN PLAY STORE)

RISA SULISTIAWATI; Mia kamayani;

Analisis Sentimen Aplikasi Maskapai Penerbangan Lion Air Menggunakan
Metode SVM dan Naïve Bayes (Studi Kasus: Ulasan Playstore)

Risa Sulistiawati

Aplikasi Lion Air merupakan sebuah aplikasi pembelian tiket pesawat yang diluncurkan pada
tanggal 21 Oktober 2014. Dapat diunduh dan digunakan di mana saja, kapan saja. Didalam
beberapa artikel berita yang melaporkan bahwa lion air merupakan maskapai penerbangan
terburuk di dunia. Namun perlu disadari aplikasi lion Air juga memiliki banyak pengguna yang
memberikan ulasan positif, negatif maupun netral dikarenakan adanya beberapa ulasan yang
dipaparkan di aplikasi Play Store. Masalah ini diteliti untuk analisis sentimen untuk mendapatkan
rating kepuasan pelanggan kepada aplikasi Lion Air dengan perolehan 2000 data. Dalam penelitian
ini, perhitungan Support Vector Machine (SVM) dan perhitungan Naive Bayes dibandingkan
dengan menggunakan 80% rasio latih dan 20% rasio uji. Dalam pertimbangan ini digunakan 795
opini positif dan 805 opini negatif yang digunakan, dimana Support Vector Machine (SVM)
dengan fitur Bigram menjadi metode paling unggul dengan presisi 99,23%, recall 83,03%, akurasi
91,75%, Skor F-1 sebesar 90,51%.
Kata Kunci: Analisis Sentimen, SVM, Naïve Bayes, Aplikasi Lion Air, Goggle Play Store
Sentiment Analysis of Lion Air Airline Application Using SVM and Naïve

Methods Bayes (Case Study: Playstore Reviews)

Risa Sulistiawati

Lion Air App is a flight ticket purchase application launched on October 21, 2014. It can be
downloaded and used anywhere, anytime. online. In several news articles reporting that Lion Air
is the world's worst airline. in the world. However, it needs to be realized that the Lion Air
application also has many users who give positive, negative and neutral reviews due to several
factors. neutral due to the existence of several reviews presented in the Play Store application.
This problem was researched for sentiment analysis to get a customer satisfaction rating for the
Lion Air application. Lion Air application with the acquisition of 2000 data. In this research,
Support Vector Machine (SVM) calculation and Naive Bayes calculation were compared using
80% training ratio and 20% test ratio. In this consideration, 795 positive opinions and 805
negative opinions were used. used, where Support Vector Machine (SVM) with Bigram features
became the most superior method with 99.23% precision. method with 99.23% precision, 83.03%
recall, 91.75% accuracy, F-1 score of 90.51%.
Keyword: Sentiment Analysis,SVM,Naïve Bayes,Lion Air Application,Google Playstore