ANALISIS SENTIMEN PENGGUNA GOPAY MENGGUNAKAN METODE LEXICON BASED DAN SUPPORT VECTOR MACHINE

Mahendrajaya, Rachmad, Buntoro, Ghulam Asrofi and Setyawan, M Bhanu (2019) ANALISIS SENTIMEN PENGGUNA GOPAY MENGGUNAKAN METODE LEXICON BASED DAN SUPPORT VECTOR MACHINE. KOMPUTEK, 3 (2). pp. 52-63. ISSN 2614-0977

[img] Text
Analisis Sentimen Pengguna Gopay Menggunakan Metode Lexicon Based dan Support Vector Machine.pdf

Download (723kB)
[img] Text
Analisis Sentimen Pengguna Gopay Menggunakan Metode Lexicon Based dan Support Vector Machine.pdf

Download (4MB)
[img] Text
Analisis Sentimen Pengguna Gopay Menggunakan Metode Lexicon Based Dan Support Vector Machine.pdf

Download (2MB)
Official URL: http://studentjournal.umpo.ac.id/index.php/kompute...

Abstract

Go-Pay is part of the Gojek application and one of the most popular finteches in Indonesia. Although the most popular, not all users have positive or even negative comments. Now users can submit various media opinions, one of which is Twitter. Twitter media has the advantage of a simple display, updated topics, open access to tweets and express opinions quickly. From a variety of comments on Twitter it takes a technique to divide into classes positive or negative opinions. This study uses prepocessing and labeling opinions into positive and negative classes with the lexicon Based method. As for the classification using the Support Vector Machine (SVM) method. The data used in the form of opinions about Go- Pay reviews from social media Twitter, amounting to 1210. The results of labeling with Lexicon Based amounted to 923 for positive and 287 for negative. While the classification of the SVM method using the Linear kernel produces 89.17% and 84.38% for the Polynomial kernel.

Item Type: Article
Uncontrolled Keywords: Sentiment Analysis, Twitter, Go-Pay, Lexicon Based, Support Vector Machine (SVM)
Subjects: T Technology > T Technology (General)
Depositing User: Library Umpo
Date Deposited: 06 Dec 2019 04:24
Last Modified: 06 Dec 2019 04:24
URI: http://eprints.umpo.ac.id/id/eprint/5115

Actions (login required)

View Item View Item