Claim Missing Document
Check
Articles

Found 1 Documents
Search

A Systematic Review of Edge Computing Architectures and Data Mining Techniques in Smart Farming Mimi Haryani Tasani; Annisa Wulan Sari
International Journal of Research and Applied Technology Vol. 6 No. 1 (2026): June 2026
Publisher : Universitas Komputer Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.34010/injuratech.v6i1.21293

Abstract

The rapid integration of the Internet of Things (IoT) in smart farming generates massive volumes of heterogeneous data, conventionally processed through centralized cloud computing. However, cloud-centric systems often suffer from high latency, bandwidth limitations, and intermittent connectivity in remote agricultural areas. Edge computing, coupled with lightweight data mining techniques, offers a transformative solution by processing data directly at the source. This study presents a Systematic Literature Review (SLR) aiming to synthesize current advancements in edge architectures and resource-efficient algorithms for precision agriculture. Following a rigorous 5-phase methodology, 38 primary studies published between 2021 and 2026 were extracted from five major academic databases (ScienceDirect, ACM, IEEE Xplore, Springer, and Web of Science). The findings indicate that the three-tier (IoT-Edge-Cloud) architecture is the most prevalent framework, effectively reducing transmission delays. For data processing on resource-constrained edge nodes, compressed tree-based models—particularly Lightweight XGBoost and Random Forest—demonstrate superior efficacy compared to computationally heavy neural networks. Nevertheless, the successful deployment of edge-AI faces substantial challenges, including heterogeneous data integration without traditional ETL pipelines and dynamic model updating against environmental concept drift. This review provides a consolidated roadmap for future research addressing these technical bottlenecks in agricultural IoT.