<![CDATA[Recommendation systems have become a crucial component in various digital applications to help users find relevant products or services based on their preferences. In the context of tourism and hospitality, recommendation systems facilitate users in selecting hotels that suit their needs and preferences. An effective approach to building a recommendation system is by using Content Based Filtering techniques. This research aims to develop a hotel recommendation model that can address the cold start problem, a situation where the recommendation system struggles to provide accurate suggestions to new users or for new items that do not yet have many interactions. By using text-mining methods, hotel descriptions and amenities are extracted into important features, which are then used by measurement methods to calculate similarity scores between features to generate relevant and accurate recommendations for users. Two similarity score measurement methods compared in this study are Cosine Similarity and RBF Kernel. The similarity score measurement were conducted using 20 hotels, where each hotel selected 10 recommended hotels with the highest similarity scores. The results showed that the RBF Kernel method outperformed with an accuracy percentage of 99.8279 %. Meanwhile, the Cosine Similarity method had a slightly lower accuracy percentage of 99,8187 %.]]>
Copyrights © 2024
