ANALISIS SENTIMEN BERBASIS ASPEK PADA APLIKASI PINJAMAN ONLINE MENGGUNAKAN LONG SHORT-TERM MEMORY

Muhammad Prasetyo

doi:10.23960/jitet.v14i3.10892

Muhammad Prasetyo
Universitas Muhammadiyah Sukabumi

DOI: https://doi.org/10.23960/jitet.v14i3.10892

Keywords Analisis Sentimen, Aspect-Based Sentiment Analysis, Long Short-Term Memory, Pinjaman Online, Google Play Store

Abstract Views (Last 12 Months)

6 Abstract Views

3 Downloads

Abstract

Abstrak. Layanan pinjaman online semakin diminati masyarakat, namun di sisi lain memunculkan berbagai keluhan terkait praktik penagihan, tingginya suku bunga, hingga penyalahgunaan data pribadi yang banyak disampaikan pengguna melalui kolom ulasan Google Play Store. Analisis sentimen konvensional belum mampu memetakan informasi secara komprehensif karena ulasan bersifat tidak terstruktur, informal, dan sering memuat lebih dari satu aspek (multi-label) dalam satu ulasan. Penelitian ini menerapkan Aspect-Based Sentiment Analysis (ABSA) untuk menganalisis persepsi pengguna terhadap aplikasi pinjaman online Easycash, Akulaku, dan Kredivo mengikuti metodologi Cross-Industry Standard Process for Data Mining (CRISP-DM). Metode Rule-Based Keyword Matching digunakan untuk mengekstraksi aspek (penagihan, bunga, privasi data) sekaligus melabeli sentimen ulasan ke dalam tiga kelas (positif, netral, negatif). Dari 30.000 data mentah hasil scraping, diperoleh 8.705 ulasan bersih yang dibagi dengan rasio 80:20 untuk melatih algoritma Long Short-Term Memory (LSTM). Hasil pengujian menunjukkan model LSTM memperoleh accuracy 82,94%, dengan rata-rata makro precision, recall, dan F1-score masing-masing 83%, serta skor ROC-AUC 0,9375. Hasil klasifikasi diimplementasikan ke dalam dashboard berbasis web yang memvisualisasikan distribusi sentimen berdasarkan aspek dan aplikasi.

Abstract. Online lending services are increasingly in demand, yet they raise complaints regarding debt collection practices, high interest rates, and personal data misuse frequently voiced through Google Play Store reviews. Conventional sentiment analysis cannot comprehensively map this information because reviews are unstructured, informal, and often multi-label. This study applies Aspect-Based Sentiment Analysis (ABSA) to examine user perceptions of the online lending applications Easycash, Akulaku, and Kredivo following the Cross-Industry Standard Process for Data Mining (CRISP-DM). A Rule-Based Keyword Matching method extracts three aspects (debt collection, interest, and data privacy) while simultaneously labeling review sentiment into three classes (positive, neutral, negative). From 30,000 raw scraped reviews, 8,705 clean reviews were obtained and split 80:20 to train a Long Short-Term Memory (LSTM) model. The model achieved 82.94% accuracy, with macro-average precision, recall, and F1-score of 83%, and a ROC-AUC score of 0.9375. The classification results were deployed into a web-based dashboard visualizing sentiment distribution by aspect and application.

Downloads

Download data is not yet available.

References

C. A. S. Hafizh Maalik Falah, “ANALISIS SENTIMEN MASYARAKAT TERHADAP PINJAMAN ONLINE DI APLIKASI X MENGGUNAKAN LONG SHORT-TERM MEMORY,” vol. 13, no. 2, 2025.

R. Sahirin, L. Lucas, and C. Qingjun, “SENTIMENT ANALYSIS ON SOCIAL MEDIA USING DATA MINING,” vol. 3, no. 1, pp. 1–11, 2025.

A. Aziz and Y. Dharta, “The Role of Social Media in Shaping Public Opinion : A Comparative Analysis of Traditional vs . Digital Media Platforms,” vol. 1, no. 6, 2024.

D. Anggraini, S. Rahmawati, and R. Kurniawan, “Natural Language Processing For Automatic Sentiment Analysis In Social Media Data,” vol. 1, no. 1, pp. 16–19, 2024.

B. A. Mustofa, W. Laksito, and Y. Saptomo, “Use of Natural Language Processing in Social Media Text Analysis,” vol. 4, no. 2, pp. 2–5, 2025.

Y. Cathy, H. Paul, D. Jörg, and W. Katerina, A systematic review of aspect ‑ based sentiment analysis : domains , methods , and trends, vol. 57, no. 11. Springer Netherlands, 2024. doi: 10.1007/s10462-024-10906-z.

R. Akbar, S. Fauzan, S. A. Wicaksono, and N. Y. Setiawan, “ANALISIS SENTIMEN BERBASIS ASPEK TERHADAP ULASAN APLIKASI QASIR MOBILE PADA GOOGLE PLAYSTORE MENGGUNAKAN METODE SUPPORT VECTOR MACHINE ( SVM ),” vol. 9, no. 6, pp. 1–10, 2025.

B. A. Putra, Y. Kristian, E. I. Setiawan, and J. Santoso, “Aspect based Sentiment Analysis Aduan Mahasiswa UMSIDA Dimasa Pandemi Menggunakan LSTM,” vol. 04, no. 01, 2022, doi: 10.52985/insyst.v4i1.229.

D. I. Af’idah, P. D. Anggraeni, M. Rizki, A. B. Setiawan, and S. F. Handayani, “Aspect-Based Sentiment Analysis for Indonesian Tourist Attraction Reviews Using Bidirectional Long Short-Term Memory,” JUITA J. Inform., vol. 11, no. 1, p. 27, 2023, doi: 10.30595/juita.v11i1.15341.

I. G. Waluyo, “Sentiment Analysis of Negative Comments on Social Media Using Long Short-Term Memory ( LSTM ) Method with TensorFlow Framework,” vol. 2, no. 7, pp. 1015–1030, 2023.

H. Zhang, “Research on Text Classification Based on LSTM-CNN,” no. January, 2024, doi: 10.1145/3708036.3708084.

A. Aljabar and A. A. A. Karim, “Analisis sentimen menggunakan algoritma lstm pada media sosial,” vol. 1, no. 3, 2022.

S. Gothane, G. V. Reddy, K. P. Kumar, D. Baswaraj, and G. P. Devi, “Sentiment Analysis in Social Media Using Deep Learning Techniques,” vol. 12, no. 3, pp. 1588–1597, 2024.

A. R. Gunawan, R. Faticha, and A. Aziza, “Sentiment Analysis Using LSTM Algorithm Regarding Grab Application Services in Indonesia,” vol. 9, no. 2, pp. 322–332, 2025.

Y. Saputri et al., “Sentiment Analysis of Trending Topics on Social Media X Using Natural Language Processing and LSTM,” vol. 9, no. 6, pp. 3034–3041, 2025.

D. D. Putri, G. F. Nama, and W. E. Sulistiono, “ANALISIS SENTIMEN KINERJA DEWAN PERWAKILAN RAKYAT ( DPR ) PADA TWITTER MENGGUNAKAN METODE NAIVE BAYES CLASSIFIER,” vol. 10, no. 1, pp. 34–40, 2022.