ANALISIS SENTIMEN ULASAN APLIKASI SAMSAT DIGITAL NASIONAL PADA GOOGLE PLAYSTORE MENGGUNAKAN ALGORITMA NAÏVE BAYES

Deni Wijaya; Faldy Irwiensyah;

Analisis Sentimen Ulasan Aplikasi Samsat Digital Nasional Pada Google Playstore Menggunakan Algoritma Naïve Bayes

Deni Wijaya

Transformasi digital telah menjadi faktor utama perubahan dalam berbagai aspek kehidupan modern, termasuk bisnis, pendidikan, dan pemerintahan. Di era transformasi digital saat ini, pemerintah berusaha meningkatkan efisiensi dan pelayanan kepada masyarakat melalui penerapan berbagai inovasi teknologi. Penerapan teknologi digital dalam layanan publik semakin meluas, termasuk di sektor pelayanan administrasi seperti Samsat Digital Nasional (SIGNAL) yang memungkinkan masyarakat dapat melakukan pembayaran pajak kendaraan secara oline melalui aplikasi SIGNAL. Evaluasi pengguna terhadap aplikasi ini dapat memberikan wawasan penting bagi penyelenggara layanan. Penelitian ini bertujuan untuk menganalisis sentimen ulasan pengguna terhadap aplikasi Samsat Digital Nasional di platform Google Playstore menggunakan algoritma Naïve Bayes. Metode ini digunakan untuk mengklasifikasikan ulasan pengguna ke dalam kategori sentimen positif dan negatif. Dari 2000 ulasan yang diambil, 1,665 ulasan dikategorikan sebagai positif dan 335 ulasan sebagai negatif setelah pelabelan manual. Proses preprocessing data menggunakan RapidMiner mencakup cleaning, transform cases, tokenizing, filter stopword, filter token by length, dan stemming. Pembobotan TF-IDF digunakan untuk memberi bobot pada setiap kata dalam dokumen. Evaluasi model Naïve Bayes menghasilkan akurasi sebesar 63.61%, dengan 307 True Positif, 74 True Negatif, 26 False Positif, dan 192 False Negatif. Precision sebesar 92.19% dan recall sebesar 61.52%. Analisis keseluruhan menunjukkan bahwa ulasan pengguna cenderung lebih positif terhadap aplikasi SIGNAL, meskipun terdapat beberapa ulasan negatif. Kesimpulan ini memberikan gambaran tentang persepsi positif pengguna terhadap aplikasi tersebut.

Kata kunci: samsat; signal; naïve bayes; rapidminer; ulasan; analisis sentimen; google playstore

Sentiment Analysis of National Digital Samsat Application Reviews on Google Playstore Using the Naïve Bayes Algorithm

Deni Wijaya

Digital transformation has become a major factor of change in various aspects of modern life, including business, education, and government. In the current era of digital transformation, the government is trying to improve efficiency and services to the community through the implementation of various technological innovations. The application of digital technology in public services is increasingly widespread, including in the administrative service sector such as the National Digital Samsat (SIGNAL) which allows people to make online vehicle tax payments through the SIGNAL application. User evaluations of this application can provide important insights for service providers. This research aims to analyze the sentiment of user reviews of the National Digital Samsat application on the Google Playstore platform using the Naïve Bayes algorithm. This method is used to classify user reviews into positive and negative sentiment categories. From 2000 reviews taken, 1,665 reviews were categorized as positive and 335 reviews as negative after manual labeling. Data preprocessing using RapidMiner includes cleaning, transform cases, tokenizing, stopword filter, token by length filter, and stemming. TF-IDF weighting is used to give weight to each word in the document. Evaluation of the Naïve Bayes model resulted in an accuracy of 63.61%, with 307 True Positives, 74 True Negatives, 26 False Positives, and 192 False Negatives. Precision was 92.19% and recall was 61.52%. The overall analysis shows that user reviews tend to be more positive towards the SIGNAL app, although there are some negative reviews. This conclusion gives an idea of users' positive perception of the app.

Keywords: samsat; signal; naïve bayes; rapidminer; reviews; sentiment analysis; google playstore