<![CDATA[Personalizing dietary plans aligned with specific medical restrictions poses a complex challenge in healthy diet planning. This study aims to develop a diet food recommendation system using a Hybrid approach that integrates Content-Based Filtering and Collaborative Filtering. The system is designed to provide recommendations that are not only personalized to user preferences but also safe regarding medical constraints such as diabetes, hypertension, obesity, and allergies. The system architecture applies a weighted average strategy with a priority weighting scheme of 0.6 for Content-Based and 0.4 for Collaborative. Performance evaluation was conducted using Leave-One-Out Cross Validation on a dataset comprising 51 users, 507 food items, and 265 interaction ratings. Test results demonstrate that the Hybrid method yields more robust performance compared to single methods, achieving a Precision of 0.7418 and Recall of 0.7500. Significantly, this approach improved Recall by 8.0% compared to pure Collaborative Filtering, proving its effectiveness in mitigating data sparsity and cold-start problems. It can be concluded that the system successfully provides relevant, promising, adaptive, and clinically safe food recommendations as a diet decision support tool, although further development in data volume and variety is required for optimal results.]]>
Copyrights © 2026
