Implementasi Metode Item-Based Collaborative Filtering dalam Rekomendasi Barang pada Aplikasi Mobile Go-BUMDes Implementation of Item-Based Collaborative Filtering Method in Product Recommendation on the Go-BUMDes Mobile Application

Abstract

This study is motivated by the low public interest in shopping at BUMDes (village-owned enterprises), primarily due to geographical constraints, highlighting the need for digital innovation to improve service accessibility. The objective of this research is to develop the GO-BUMDes mobile application as a platform for product ordering and recommendation in Cau Belayu Village. The application employs an Item-Based Collaborative Filtering method to provide product recommendations based on item similarity. The development process followed the prototype methodology, while system testing involved white box and black box techniques, accuracy evaluation using MAE (Mean Absolute Error), and user experience assessment through UMUX (Usability Metric for User Experience). Test results showed an MAE value of 0.258, indicating a relatively high prediction accuracy, and a UMUX score of 85.78, reflecting excellent user comfort and satisfaction. The study concludes that GO-BUMDes has the potential to enhance access and facilitate digital transactions at BUMDes, while encouraging community participation in a technology-driven village economy. The practical implications of this research contribute to strengthening digital transformation in the local economic sector, particularly in rural areas.