Analisis Sentimen Ulasan Pengguna Aplikasi Pospay Biru Menggunakan Kombinasi TF-IDF dan Support Vector Machine

Ade Mulyadi (Program Studi Diploma IV Teknik Informatika, Fakultas Sekolah Vokasi, Universitas Logistik dan Bisnis Internasional, Bandung)

Sinergi : Jurnal Ilmiah Multidisiplin
Vol. 2 No. 8 (2026): Sinergi: Jurnal Ilmiah Multidisiplin

Transformasi layanan keuangan digital mendorong meningkatnya penggunaan aplikasi Pospay Biru sebagai media transaksi elektronik milik PT Pos Indonesia. Seiring bertambahnya jumlah pengguna, ulasan pada Google Play Store menjadi sumber informasi penting untuk mengevaluasi kualitas layanan dan tingkat kepuasan pengguna. Namun, tingginya volume ulasan menyebabkan analisis secara manual menjadi tidak efisien dan rentan terhadap subjektivitas. Penelitian ini bertujuan menganalisis sentimen ulasan pengguna Pospay Biru menggunakan kombinasi Term Frequency–Inverse Document Frequency (TF-IDF) sebagai metode ekstraksi fitur dan Support Vector Machine (SVM) sebagai algoritma klasifikasi. Data penelitian diperoleh melalui proses web scraping Google Play Store yang menghasilkan 25.391 ulasan, kemudian melalui tahapan seleksi dan preprocessing sehingga diperoleh 5.971 ulasan yang layak dianalisis. Tahapan preprocessing meliputi cleaning, case folding, tokenizing, stopword removal, dan stemming. Selanjutnya, setiap ulasan direpresentasikan dalam bentuk vektor TF-IDF dan diklasifikasikan menggunakan SVM dengan linear kernel. Evaluasi model dilakukan menggunakan confusion matrix, accuracy, precision, recall, dan F1-score. Hasil penelitian menunjukkan bahwa model menghasilkan accuracy sebesar 89,13%, precision 83,93%, recall 84,62%, dan F1-score 84,27%. Temuan ini menunjukkan bahwa kombinasi TF-IDF dan SVM mampu mengklasifikasikan sentimen ulasan pengguna secara efektif serta dapat dimanfaatkan sebagai pendekatan otomatis untuk mendukung evaluasi kualitas layanan aplikasi Pospay Biru berdasarkan umpan balik pengguna. Kata Kunci: Analisis Sentimen, Pospay Biru, Text Mining, TF-IDF, Support Vector Machine. Abstract The rapid growth of digital financial services has increased the use of Pospay Biru, a digital payment application developed by PT Pos Indonesia. As the number of users continues to grow, reviews posted on Google Play Store provide valuable information for evaluating service quality and user satisfaction. However, the large volume of reviews makes manual analysis inefficient and prone to subjective interpretation. This study aims to analyze user sentiment toward the Pospay Biru application by combining Term Frequency–Inverse Document Frequency (TF-IDF) for feature extraction and Support Vector Machine (SVM) for sentiment classification. The dataset was collected through Google Play Store web scraping, producing 25,391 raw reviews. After data selection and preprocessing, 5,971 valid reviews were retained for analysis. The preprocessing stage consisted of cleaning, case folding, tokenization, stopword removal, and stemming. The processed reviews were transformed into TF-IDF feature vectors and classified using a linear-kernel SVM model. Model performance was evaluated using a confusion matrix, accuracy, precision, recall, and F1-score. The experimental results achieved an accuracy of 89.13%, precision of 83.93%, recall of 84.62%, and an F1-score of 84.27%. These findings demonstrate that the combination of TF-IDF and SVM provides effective sentiment classification for Indonesian-language application reviews and can serve as an automated approach to support service quality evaluation and decision-making for improving the Pospay Biru application. Keywords: Sentiment Analysis, Pospay Biru, Text Mining, TF-IDF, Support Vector Machine.