Analisis Sentimen Berdasarkan Topik Ulasan Aplikasi Akseleran di Google Play Store Menggunakan Naïve Bayes Classifier dan Latent Dirichlet Allocation

Septiadi, Faniya Mahesty (2026) Analisis Sentimen Berdasarkan Topik Ulasan Aplikasi Akseleran di Google Play Store Menggunakan Naïve Bayes Classifier dan Latent Dirichlet Allocation. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2043221033-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (1MB) | Request a copy

Abstract

Financial Technology (fintech) di Indonesia terus mengalami pertumbuhan, salah satunya melalui layanan peer-to-peer (P2P) lending. Akseleran (PT Akselerasi Usaha Indonesia) merupakan salah satu platform P2P lending yang hingga Januari 2025 telah menyalurkan pendanaan sebesar Rp12,42 triliun dan memperoleh ribuan ulasan dengan rating rata-rata 3,5 di Google Play Store. Ulasan tersebut mencerminkan opini pengguna mengenai berbagai aspek layanan, namun jumlahnya yang besar dan berbentuk teks tidak terstruktur membuat analisis manual menjadi tidak efisien. Penelitian ini menerapkan pendekatan text mining dengan mengombinasikan Latent Dirichlet Allocation (LDA) dan Naïve Bayes Classifier (NBC). LDA digunakan untuk mengidentifikasi topik-topik utama dalam 6.582 ulasan, menghasilkan dua tema utama yaitu Penggunaan Aplikasi dan Aktivitas Usaha; dan Proses Verifikasi Data dan Pengajuan Pinjaman, dengan nilai coherence score optimal 0,5752. Selanjutnya, NBC diterapkan pada setiap topik untuk mengklasifikasikan polaritas ulasan ke dalam kategori positif atau negatif berdasarkan representasi Term Frequency-Inverse Document Frequency (TF-IDF). Hasil klasifikasi menunjukkan bahwa pada topik pertama, sentimen positif mendominasi sebesar 97,0% dengan akurasi model 97,88%. Sementara pada topik kedua, distribusi sentimen hampir berimbang yaitu 51,5% positif dan 48,5% negatif dengan akurasi model 87,86%. Temuan ini mengindikasikan bahwa pengguna sangat puas dengan kemudahan penggunaan aplikasi dan manfaatnya bagi usaha, namun memiliki keluhan signifikan terkait proses verifikasi dan pengajuan pinjaman. Penelitian ini memberikan rekomendasi bahwa prioritas perbaikan layanan sebaiknya difokuskan pada penyederhanaan dan percepatan proses verifikasi data serta pengajuan pinjaman.
=================================================================================================================================
Financial Technology (fintech) in Indonesia continues to grow, one of which is through peer-to-peer (P2P) lending services. Akseleran (PT Akselerasi Usaha Indonesia) is one of the P2P lending platforms that, as of January 2025, has disbursed funding amounting to IDR 12.42 trillion and received thousands of reviews with an average rating of 3.5 on the Google Play Store. These reviews reflect users’ opinions regarding various aspects of the service; however, their large volume and unstructured textual form make manual analysis inefficient. This study employs a text mining approach by integrating Latent Dirichlet Allocation (LDA) and the Naïve Bayes Classifier (NBC). LDA is used to identify key topics within 6,582 reviews and produces two main themes: (1) Application Usage and Business Activities, and (2) Data Verification and Loan Submission Processes, with an optimal coherence score of 0.5752. Subsequently, NBC is applied to each topic to classify review polarity into positive or negative categories based on Term Frequency–Inverse Document Frequency (TF-IDF) representations. The classification results show that for the first topic, positive sentiment dominates at 97.0%, with a model accuracy of 97.88%. Meanwhile, for the second topic, sentiment distribution is relatively balanced, consisting of 51.5% positive and 48.5% negative, with a model accuracy of 87.86%. These findings indicate that users are highly satisfied with the ease of application use and its benefits for business activities, but express significant concerns regarding the verification and loan submission processes. This study suggests that service improvement efforts should prioritize simplifying and accelerating the data verification and loan submission processes.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Akseleran, Naïve Bayes Classifier, Latent Dirichlet Allocation, Pemodelan Topik, Analisis Sentimen Akseleran, Naïve Bayes Classifier, Latent Dirichlet Allocation, Sentiment Analysis, Topic Modeling
Subjects:	Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines.
Divisions:	Faculty of Vocational > 49501-Business Statistics
Depositing User:	Faniya Mahesty Septiadi
Date Deposited:	10 Feb 2026 03:33
Last Modified:	10 Feb 2026 03:33
URI:	http://repository.its.ac.id/id/eprint/132307

Actions (login required)

View Item