Abstract
Abstract: SIAPkerja is a digital employment service platform organized by the Ministry of Manpower of the Republic of Indonesia to facilitate interactions between job seekers and employers. The increasing number of application users has resulted in a higher volume of reviews on the Google Play Store platform, reflecting diverse user experiences and satisfaction levels with the provided services. This study analyzes the sentiment of SIAPkerja user reviews, with 2,000 review data points collected through web scraping from the Google Play Store during the period from October 2022 to June 2026. The data were processed through cleaning, case folding, tokenization, normalization, stopword removal, and stemming, and then represented using TF-IDF weighting. K-Means was applied to cluster public opinions based on analytical patterns and sentiment as the basis for initial label determination, while SVM and Random Forest were used to classify these opinions into positive and negative sentiment categories. SVM achieved an accuracy of 99.63%, which was slightly higher than Random Forest. These results indicate that sentiment analysis based on text mining can be used to evaluate the quality of SIAPkerja services from the perspective of users and can serve as a consideration in formulating digital employment service policies that are more responsive to user needs. Keywords: SIAPkerja, Google Play Store, Sentiment Analysis, Support Vector Machine, Random Forest. Abstrak: SIAPkerja merupakan platform layanan ketenagakerjaan digital yang diselenggarakan oleh Kementerian Ketenagakerjaan Republik Indonesia untuk memfasilitasi interaksi antara pencari kerja dan pemberi kerja. Peningkatan jumlah pengguna aplikasi tersebut berimplikasi pada bertambahnya volume ulasan pada platform Google Play Store yang merefleksikan variasi pengalaman serta tingkat kepuasan pengguna terhadap layanan yang disediakan. Penelitian ini menganalisis sentimen ulasan pengguna SIAPkerja sebanyak 2.000 data ulasan dikumpulkan melalui web scraping dari Google Play Store pada periode Oktober 2022 hingga Juni 2026. Data diproses melalui tahap cleaning, case folding, tokenisasi, normalisasi, stopword removal, dan stemming, kemudian direpresentasikan dengan pembobotan TF-IDF. K-Means diterapkan untuk mengelompokkan opini masyarakat berdasarkan pola analisis dan sentimen sebagai dasar penentuan label awal, sementara SVM dan Random Forest digunakan untuk mengklasifikasikan opini tersebut ke dalam kategori sentimen positif dan negatif. SVM menghasilkan akurasi 99,63%, sedikit lebih tinggi dibanding Random Forest. Hasil ini menunjukkan bahwa analisis sentimen berbasis text mining dapat digunakan untuk mengevaluasi kualitas layanan SIAPkerja berdasarkan persepsi pengguna, dan dapat menjadi bahan pertimbangan dalam penyusunan kebijakan layanan ketenagakerjaan digital yang lebih responsif terhadap kebutuhan pengguna. Kata kunci: SIAPkerja, Google Play Store, Analisis Sentimen, Support Vector Machine, Random Forest.
