Smartwatch Selection Recommendation System Using the K-Nearest Neighbor (KNN) Algorithm with Dynamic Dataset Optimization

This research aimed to develop a smartwatch recommendation system using the K-Nearest Neighbor (KNN) algorithm with dynamic dataset optimization. By employing a dynamic dataset, the accuracy of KNN calculations was enhanced. The dataset, stored in CSV format, was filtered based on user preferences when searching for a smartwatch, generating a dynamic dataset tailored to individual needs. The research involved 35 respondents to evaluate the precision and feasibility of the application. Results showed that 25.7% of respondents found the application highly relevant to their preferences, 31.4% relevant, and 31.4% somewhat relevant. User satisfaction levels indicated that 34.3% were very satisfied, 34.3% satisfied, and 20% somewhat satisfied, highlighting the application’s effectiveness in meeting user expectations.