ANALISIS SENTIMEN ULASAN PENGGUNA APLIKASI WONDR MENGGUNAKAN HYBRID RFSVM MODEL

RIFA KHAIRUNNISA PRATIWI, . (2025) ANALISIS SENTIMEN ULASAN PENGGUNA APLIKASI WONDR MENGGUNAKAN HYBRID RFSVM MODEL. Sarjana thesis, UNIVERSITAS NEGERI JAKARTA.

	Text COVER.pdf Download (1MB)
	Text BAB I.pdf Download (354kB)
	Text BAB II.pdf Restricted to Registered users only Download (1MB) | Request a copy
	Text BAB III.pdf Restricted to Registered users only Download (369kB) | Request a copy
	Text BAB IV.pdf Restricted to Registered users only Download (1MB) | Request a copy
	Text BAB V.pdf Restricted to Registered users only Download (224kB) | Request a copy
	Text DAFTAR PUSTAKA.pdf Download (230kB)
	Text LAMPIRAN.pdf Restricted to Registered users only Download (4MB) | Request a copy

Abstract

Wondr merupakan platform mobile banking BNI yang membantu nasabah mengelola keuangan dengan efisien. Untuk menilai kualitas layanan dan kepuasan pengguna, dilakukan analisis sentimen terhadap ulasan pengguna. Penelitian ini menerapkan metode hybrid yang menggabungkan algoritma Random Forest (RF) dan Support Vector Machine (SVM), disebut hybrid RFSVM. Tujuan penelitian ini adalah menjelaskan proses penggabungan model serta mengukur kinerjanya dalam klasifikasi ulasan aplikasi Wondr. Data diambil dari Google Play Store pada 8 Maret hingga 30 April 2025. Hasil menunjukkan akurasi 95% dengan nilai presisi untuk kelas negatif, netral, dan positif masing-masing 82%, 94%, dan 97%. Nilai recall untuk kelas negatif, positif, dan netral berturut-turut sebesar 85%, 99%, dan 58%, sedangkan F1-score masing-masing 84%, 98%, dan 72%. Berdasarkan hasil klasifikasi diperoleh 5.419 ulasan positif, 749 ulasan negatif, dan 137 ulasan netral. Secara keseluruhan, mayoritas pengguna memiliki persepsi positif terhadap aplikasiWondr, meskipun masih terdapat ulasan negatif dan netral yang menunjukkan perlunya peningkatan pada beberapa aspek layanan. ***** Wondr is BNI’s mobile banking platform that helps customers manage their finances efficiently. To assess service quality and user satisfaction, sentiment analysis was conducted on user reviews. This study applied a hybrid method that combines the Random Forest (RF) and Support Vector Machine (SVM) algorithms, called hybrid RFSVM. The purpose of this study is to explain the model combination process and measure its performance in classifying Wondr app reviews. Data was collected from the Google Play Store from March 8 to April 30, 2025. The results show an accuracy of 95% with precision values for negative, neutral, and positive classes of 82%, 94%, and 97%, respectively. The recall values for the negative, positive, and neutral classes were 85%, 99%, and 58%, respectively, while the F1-scores were 84%, 98%, and 72%, respectively. Based on the classification results, there were 5,419 positive reviews, 749 negative reviews, and 137 neutral reviews. Overall, the majority of users had a positive perception of the Wondr app, although there were still negative and neutral reviews that indicated the need for improvement in several aspects of the service.

Item Type:	Thesis (Sarjana)
Additional Information:	1). Qorry Meidianingsih, S.Si., M.Si. ; 2). Devi Eka Wardani Meganingtyas, S.Pd., M.Si.
Subjects:	Sains > Matematika
Divisions:	FMIPA > S1 Matematika
Depositing User:	Users 32783 not found.
Date Deposited:	05 Feb 2026 08:06
Last Modified:	05 Feb 2026 08:06
URI:	http://repository.unj.ac.id/id/eprint/65006

Actions (login required)

View Item