ANALISIS SENTIMEN TANGGAPAN PENGGUNA APLIKASI BALE BY BTN MENGGUNAKAN METODE SUPPORT VECTOR MACHINE (SVM)

Ahmat Setiawan; Firman Noor Hasan;

Analisis Sentimen Tanggapan Pengguna Aplikasi Bale By Btn Menggunakan

Metode Support Vector Machine (Svm)

Ahmat Setiawan

Dalam era digital yang terus berkembang, analisis sentimen terhadap ulasan pengguna menjadi alat
penting untuk mengevaluasi kualitas aplikasi mobile banking. Penelitian ini bermaksud untuk
mengetahui sentimen pengguna terhadap aplikasi bale by BTN yang diluncurkan pada Februari 2025
sebagai penyempurna dari aplikasi BTN Mobile. Metode yang digunakan meliputi scraping data
ulasan dari Google Play Store, preprocessing teks (cleaning, case folding, normalization, tokenisasi,
stopword removal, dan stemming), pelabelan berdasarkan kamus lexicon-based approach, serta
pembangunan model klasifikasi yang digunakan adalah algoritma Support Vector Machine (SVM)
dengan TF-IDF vectorization. Dari 2.000 data awal, diperoleh 1.767 data valid yang dianalisis. Hasil
menunjukkan bahwa model SVM mencapai akurasi sebesar 73,16%, dari 354 data testing dengan
distribusi sentimen: positif (52,57%), dan negatif (47,43%). Model menunjukkan performa terbaik
dalam mengklasifikasi sentimen Positif dengan precision 0.73, recall 0.80, dan F1-score 0,77 pada
194 data sedangkan pada sentimen negatif, model menunjukan hasil yang baik dengan precision
0.73, recall 0.65, dan F1-score 0.69 pada 160 data. Penelitian ini mengidentifikasi pola sentimen dan
masalah dominan pada aplikasi super app bale by BTN pada tahap awal peluncuran, memberikan
insight strategis untuk perbaikan prioritas dan metodologi analisis sentimen yang dapat diterapkan
pada aplikasi mobile banking serupa.
Kata kunci: analisis sentimen, bale by btn, support vector machine, google playstore
Application of Naïve Bayes Algorithm for Sentiment Analysis of Public Opinion

on Maxim Service Usage
Ahmat Setiawan

In the rapidly evolving digital era, sentiment analysis of user reviews has emerged as a crucial
instrument for evaluating mobile banking application quality. This study aims to investigate user
sentiment toward the Bale by BTN application, which was launched in February 2025 as an
enhanced successor to the BTN Mobile application. The methodology encompasses data scraping
of reviews from Google Play Store, text preprocessing (cleaning, case folding, normalization,
tokenization, stopword removal, and stemming), labeling through lexicon-based approach, and

development of a classification model utilizing Support Vector Machine (SVM) algorithm with TF-
IDF vectorization. From an initial dataset of 2,000 entries, 1,767 valid data points were obtained

for analysis. The results demonstrate that the SVM model achieved an accuracy of 73.16% on 354
testing data with sentiment distribution of positive (52.57%) and negative (47.43%). The model
exhibited optimal performance in classifying positive sentiment with precision of 0.73, recall of 0.80,
and F1-score of 0.77 across 194 data points, while for negative sentiment, the model demonstrated
satisfactory results with precision of 0.73, recall of 0.65, and F1-score of 0.69 across 160 data
points. This research identifies sentiment patterns and predominant issues within the Bale by BTN
super app during its initial launch phase, providing strategic insights for priority improvements and
sentiment analysis methodology applicable to similar mobile banking applications.data.
Keywords: sentiment analysis, bale by btn, support vector machine, google playstore