<![CDATA[Sleman Regency, Special Region of Yogyakarta, is home to 209 barbershops distributed across 13 of its 17 districts. The wide range of available barbershops can make it difficult for consumers to identify options that match their preferences. This study develops a barbershop recommendation system using Content-Based Filtering (CBF) with Cosine Similarity. The dataset was collected through web scraping from Google Maps, incorporating features such as barbershop name, district, rating, number of reviews, and service category. A feature-weighting experiment was conducted to determine the optimal weight for the district feature by testing 1×, 2×, 3×, and 4× weighting schemes. The results showed that a 3.0× weight achieved the best performance. Evaluation on 78 samples using Precision@N, Recall@N, and F1-score produced a Recall@10 of 0.3205 and an F1-score@10 of 0.1307. The Precision@10 of 0.0821 was relatively low due to the homogeneous rating distribution in the dataset (4.3–5.0), which caused most barbershops within the same district to satisfy the ground-truth criteria. Nevertheless, the results demonstrate that the proposed system can provide relevant barbershop recommendations in Sleman Regency, particularly in terms of recall-based coverage of relevant items. The relatively low precision indicates the need for additional discriminative features, such as price, distance, specific services, facilities, or user preferences, to improve recommendation quality in future research.]]>
Copyrights © 2026
