AI-DRIVEN AGRICULTURE FRAMEWORK USING ENSEMBLE ML WEBGIS AND RAG SYSTEM

Authors

  • Macalisang, DIG. Southwestern University PHINMA
  • Magdadaro, JJM. Southwestern University PHINMA
  • Say, RAL Southwestern University PHINMA
  • Ilustrisimo, AJB. Southwestern University PHINMA

Abstract

The agricultural advisory process in the Philippines is constrained by manual soil analysis, static regulatory mapping, and systemic administrative latency. This study evaluates the operational bottlenecks experienced by the Department of Agriculture, Region 7, and proposes an Internet of Things (IoT) and Machine Learning (ML) framework to advance precision agriculture and modernize extension services. Utilizing a mixed-methods approach across 41 stakeholders in Dalaguete, Cebu, the baseline assessment revealed critical limitations, including a lack of in-situ diagnostic hardware and reliance on error-prone visual slope estimations. These manual workflows contribute to a 30-day turnaround time for official advisories. To resolve these inefficiencies, this study introduces an integrated architecture consisting of three conceptual domains: IoT perception and tiered data acquisition, a cloud-based web system engine, and a multi-tiered Artificial Intelligence (AI) synthesis pipeline. The framework leverages ensemble machine learning for crop suitability forecasting, Web Geographic Information System (WebGIS)-based regulatory compliance enforcement, and Retrieval-Augmented Generation (RAG) for localized advisory synthesis. By centralizing disparate data streams into a unified processing pipeline, the architecture effectively bypasses administrative bottlenecks, providing a scalable, data-driven, and legally compliant solution for agricultural extension efforts.

Downloads

Published

2026-11-17