<![CDATA[The South Kalimantan Tax Info application service can provide information about the amount of tax, when you have to pay tax, the due date, and so on. However, the South Kalimantan Tax Info application received many negative reviews from users. So it is necessary to analyze user sentiment using computational techniques. In this context, sentiment analysis is applied using the Naïve Bayes method to user reviews, assisting the Regional Revenue Agency in understanding perceptions and enhancing service quality. The literature review encompasses similar studies that employ the Naïve Bayes algorithm for sentiment analysis in e-government applications. The research methodology involves collecting review data from the Google Play Store through web scraping, labeling based on ratings, and pre-processing. The results of sentiment analysis, utilizing the confusion matrix, demonstrate the highest accuracy of 92% with a 10:90 data split. This study contributes to users' comprehension of public service applications, facilitating continuous improvement.Keywords: Sentiment analysis; Confusion matrix; Text mining; Emoticon AbstrakLayanan aplikasi Info Pajak Kalsel dapat memberikan informasi tentang besaran pajak, waktu harus bayar pajak, jatuh tempo dan lain-lain. Namun, aplikasi Info Pajak Kalsel menerima banyak ulasan negatif dari pengguna. Maka diperlukan analisis sentimen pengguna menggunakan teknik komputasi. Dalam konteks ini, analisis sentimen diterapkan menggunakan metode Naïve Bayes pada ulasan pengguna, membantu Badan Pendapatan Daerah memahami persepsi dan meningkatkan kualitas layanan. Studi literatur mencakup penelitian sejenis yang menggunakan algoritma Naïve Bayes untuk analisis sentimen pada aplikasi e-Government. Metodologi penelitian melibatkan pengumpulan data ulasan dari Google Playstore melalui web scraping, pemberian label berdasarkan rating, dan proses pre-processing. Hasil analisis sentimen menggunakan confusion matrix menunjukkan akurasi tertinggi sebesar 92% pada pembagian data 10:90. Studi ini memberikan kontribusi pada pemahaman pengguna terhadap aplikasi pelayanan publik, memungkinkan perbaikan berkelanjutan.Kata kunci: Analisis sentimen; Matriks konfusi; Text mining; Emotikon]]>
Copyrights © 2024
