<![CDATA[This study designs and develops a decentralized offline-first knowledge management system to support an adaptive learning ecosystem that integrates the ethnomathematics of Kodi woven fabric in Southwest Sumba Regency, East Nusa Tenggara. The system is designed to address extreme digital infrastructure limitations in 3T (frontier, outermost, disadvantaged) regions by implementing Progressive Web Application (PWA), WebRTC-based peer-to-peer (P2P) technology for internet-independent data synchronization, and Bayesian Knowledge Tracing (BKT) adaptive learning algorithms running on-device through TensorFlow.js. The study employs a design-based research approach involving traditional weavers, customary leaders, mathematics teachers, students, and school principals from four districts of the Kodi Region as test subjects. The system implements the SECI model in digital architecture through social collaboration platforms, multimedia-based tacit knowledge externalization systems with speech-to-text and computer vision technology, Neo4j knowledge graph, and adaptive practice systems with spaced repetition algorithms. Test results show a System Usability Scale (SUS) score of 82.4 ("Excellent"), a 37.8% increase in learning effectiveness (pre-test 58.2 ? post-test 80.2, t(45)=8.94, p<0.01), and 94.7% offline synchronization success under limited network conditions. The system contributes to the development of inclusive educational technology for 3T regions and the preservation of indigenous knowledge through human-centered design that empowers indigenous communities as equal partners.]]>
Copyrights © 2026
