Penerapan Content-Based Filtering dan K-Nearest Neighbor pada Sistem Rekomendasi Wisata Probolinggo

Syamsul Arifin (Institut Teknologi Dan Bisnis Asia Malang)
Rina Dewi Indahsari (Institut Teknologi Dan Bisnis Asia Malang)

Progresif: Jurnal Ilmiah Komputer
Vol 22, No 1 (2026): Januari

Probolinggo Regency has a wide range of tourism destinations; however, the limited dissemination of tourism information makes it difficult for tourists to determine suitable travel destinations. This study aims to develop a tourism destination recommendation system to assist users in obtaining recommendations based on their preferences. The system is developed using the Content-Based Filtering (CBF) method with TF-IDF weighting and cosine similarity, as well as the K-Nearest Neighbor (KNN) algorithm based on cosine distance. The system process includes tourism data collection, text feature extraction and weighting, similarity calculation, and the presentation of main destination recommendations and similar destinations. System evaluation using precision and recall metrics shows that the Content-Based Filtering (CBF) and K-Nearest Neighbor (KNN) methods achieve precision values of 93.75% and 96.00%, respectively, with both methods obtaining a recall value of 100%, indicating that the system is able to provide relevant tourism destination recommendations.Keywords: Recommendation System; Probolinggo Tourism; Content-Based Filtering; TF-IDF;  K-Nearest Neighbor; Cosine Similarity. AbstrakKabupaten Probolinggo memiliki potensi destinasi wisata yang beragam, namun penyebaran informasi wisata yang belum optimal menyulitkan wisatawan dalam menentukan tujuan perjalanan. Penelitian ini bertujuan mengembangkan sistem rekomendasi destinasi wisata untuk membantu pengguna memperoleh rekomendasi sesuai preferensi. Sistem dikembangkan menggunakan metode Content-Based Filtering (CBF) dengan pembobotan TF-IDF dan cosine similarity serta algoritma K-Nearest Neighbor (KNN) berbasis cosine distance. Proses sistem meliputi pengumpulan data destinasi wisata, ekstraksi dan pembobotan fitur teks, perhitungan tingkat kemiripan, serta penyajian rekomendasi destinasi utama dan destinasi serupa. Pengujian sistem menggunakan metrik precision dan recall menunjukkan bahwa metode content-based filtering (CBF) dan K-Nearest Neighbor (KNN) menghasilkan nilai precision masing-masing sebesar 93,75% dan 96,00%, dengan nilai recall keduanya mencapai 100%, sehingga sistem mampu memberikan rekomendasi destinasi wisata yang relevan.Kata kunci: Sistem Rekomendasi; Pariwisata Probolinggo; Content-Based Filtering; TF-IDF;  K-Nearest Neighbor; Cosine Similarity.