ANALISIS SENTIMEN ULASAN PENGGUNA APLIKASI SEABANK MENGGUNAKAN SUPPORT VECTOR MACHINE DAN NAÏVE BAYES

Nada Adela, Cindy and Sri Karnila, Sri (2024) ANALISIS SENTIMEN ULASAN PENGGUNA APLIKASI SEABANK MENGGUNAKAN SUPPORT VECTOR MACHINE DAN NAÏVE BAYES. Skripsi thesis, Institut Informatika Dan Bisnis Darmajaya.

[img] Text
COVER.pdf
Download (255kB)
[img] Text
ABSTRAK.pdf
Download (159kB)
[img] Text
PERNYATAAN.pdf
Download (95kB)
[img] Text
PENGESAHAN.pdf
Download (103kB)
[img] Text
DAFTAR ISI.pdf
Download (241kB)
[img] Text
BAB 1.pdf
Download (226kB)
[img] Text
BAB 2.pdf
Download (378kB)
[img] Text
BAB 3.pdf
Download (4MB)
[img] Text
BAB 4.pdf
Download (5MB)
[img] Text
BAB 5.pdf
Download (169kB)
[img] Text
DAFTAR PUSTAKA.pdf
Download (222kB)

Abstract

Seabank is a digital banking application that offers various services such as transfers, savings and investment transactions to users. The implementation of Seabank faces a large number of user reviews related to the service. So far, the review data data has not been managed properly, for that sentiment analysis is needed in understanding positive, neutral, and negative sentiments. This research not only analyzed but also identified deeper sentiments on customer satisfaction, weaknesses of the app, and determined areas of improvement needed. The method used with two classification algorithms, namely Support Vector Machine and Naïve Bayes, the data used was 3789 review data, 80% of which was divided by 80%. Data used was 3789 review data, 80% training and 20% testing. Labeling results 438 positive sentiments, 1379 neutral sentiments, and 77 negative sentiments, while the labeling results by linguists showed a slightly different number, namely 1100 positive sentiments, 308 neutral sentiments, and 486 negative sentiments. Accuracy analysis using confusion matrix showed that the Support Vector Machine algorithm has the highest accuracy value of 63%, while the Gaussian Naïve Bayes algorithm has the lowest value of 30%. Based on these results, it showed that the Support Vector Machine model was more effective in classifying user review sentiment Seabank app compared to the Naïve Bayes model. An important contribution in understanding user sentiment trends towards the Seabank app as well as the use of classification algorithms for sentiment analysis. Keywords: Seabank, Sentiment Analysis, Support Vector Machine, Naïve Bayes

Item Type: Thesis (Skripsi)
Subjects: Ilmu Komputer
eSkripsi
Divisions: Skripsi/TA & PKPM/KP - Fakultas Ilmu Komputer > Sistem Informasi
Depositing User: Cindy Nada Adela
Date Deposited: 05 Aug 2024 07:28
Last Modified: 05 Aug 2024 07:28
URI: http://repo.darmajaya.ac.id/id/eprint/16947

Actions (login required)

View Item View Item