Sistem Rekomendasi Destinasi Wisata Di Bali Menggunakan Metode Hybrid Collaborative Filtering dan Content-Based Filtering

Chairunisa, Sindy (2023) Sistem Rekomendasi Destinasi Wisata Di Bali Menggunakan Metode Hybrid Collaborative Filtering dan Content-Based Filtering. Other thesis, Institut Teknologi Sepuluh Nopember.

[thumbnail of 06211940000105-Undergraduate_Thesis.pdf] Text
06211940000105-Undergraduate_Thesis.pdf - Accepted Version
Restricted to Repository staff only until 1 October 2025.

Download (2MB) | Request a copy

Abstract

Salah satu sektor yang yang menjadi salah satu satu sumber bagi penerimaan devisa adalah sektor pariwisata. Bali merupakan salah satu ikon pariwisata Indonesia di mata dunia. Bali memiliki beragam destinasi wisata, mulai wisata alam, wisata budaya dan kesenian, wisata sejarah, dan masih banyak lagi. Hal ini menyebabkan kesulitan bagi wisatawan untuk menemukan informasi mengenai destinasi wisata yang sesuai dengan kebutuhan mereka. Sistem rekomendasi merupakan salah satu solusi untuk memberikan rekomendasi ketika pengguna dihadapkan dengan jumlah informasi destinasi wisata yang besar. Metode yang biasa digunakan adalah collaborative filtering dan contend-based filtering. Content-based filtering memberikan rekomendasi berdasarkan kemiripan konten. Sedangkan collaborative filtering memprediksi rating atau penilaian wisatawan terhadap destinasi wisata. Namun kelemahan yang dimiliki adalah sparsity data. Oleh karena pada penelitian ini digunakan metode digunakan hybrid collaborative filtering dan contend-based filtering untuk dengan memanfaatkan content-based filtering dalam mengisi rating kosong yang selanjutnya dengan rating yang telah diisi, menjadi input dalam metode collaborative filtering yang harapannya mampu mengatasi sparsity data. Berdasarkan hasil eksperimen yang dilakukan, didapatkan hasil bahwa hybrid filtering terbukti cukup akurat dan lebih baik dalam memberikan rekomendasi dengan hasil eksperimen yang menghasilkan RMSE terendah adalah 0,6284. RMSE tersebut dibandingkan dengan metode content-based filtering yang menghasilkan RMSE sebesar 2,8249 dan metode collaborative filtering yang menghasilkan RMSE sebesar 1,0414. Dari metode hybrid filtering didapatkan accuracy sebesar 0,7694 atau sebesar 76,94%, precision sebesar 0,803 atau 80,3%, recall sebesar 0,9296 atau sebesar 92,96%, dan F1-Score sebesar 0,8617 atau sebesar 86,17%. Metode dengan RMSE terkecil yaitu hybrid filtering, digunakan untuk memberikan rekomendasi sejumlah 10 destinasi wisata berserta informasi destinasi wisata mulai dari deskripsi, harga hari libur, harga hari kerja, kategori, rating dari seluruh wisatawan lain, dan alamat dari wisata tersebut.
======================================================================================================================================
One sector that serves as a significant source of foreign exchange earnings is the tourism sector. Bali is one of Indonesia's tourism icons in the eyes of the world. Bali offers a diverse range of tourist destinations, including natural attractions, cultural and artistic experiences, historical sites, and much more. This variety can make it challenging for travelers to find information about destinations that suit their preferences. A recommendation system is one of the solutions to provide personalized recommendations when users are faced with a large amount of tourism information. Commonly used methods are collaborative filtering and content-based filtering. Content-Based Filtering provides recommendations based on content similarity, while collaborative filtering predicts ratings or evaluations by tourists for tourist destinations. However, one of the weaknesses is sparsity data. Therefore, in this study, a hybrid approach using collaborative filtering and content-based filtering is employed, utilizing content-based filtering to fill in missing ratings, followed by collaborative filtering to address sparsity data. The experimental results show that the hybrid filtering method proves to be accurate and superior in providing recommendations, with the lowest RMSE obtained at 0,6284. This is compared to content-based filtering, which resulted in an RMSE of 2,8249, and collaborative filtering, which produced an RMSE of 1,0414. The hybrid filtering method achieves an accuracy of 0,7694 or 76,94%, precision of 0,803 or 80,3%, recall of 0,9296 or 92,96%, and F1-Score of 0,8617 or 86,17%. The hybrid filtering method with the lowest RMSE is then used to provide recommendations for ten tourist destinations, along with detailed information about each destination, including descriptions, holiday and weekday prices, categories, ratings from other tourists, and addresses.

Item Type: Thesis (Other)
Uncontrolled Keywords: Collaborative Filtering, Contend-Based Filtering, Hybrid Filtering, Pariwisata Bali, Sistem Rekomendasi, Bali Tourism, Collaborative Filtering, Contend-Based Filtering, Hybrid Filtering, Recommendation System
Subjects: H Social Sciences > HA Statistics
Q Science > QA Mathematics > QA76.9.I58 Recommender systems (Information filtering)
Divisions: Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User: Sindy Chairunisa
Date Deposited: 14 Aug 2023 08:27
Last Modified: 14 Aug 2023 08:29
URI: http://repository.its.ac.id/id/eprint/104689

Actions (login required)

View Item View Item