Monitoring and Prediction of Hass Avocado Conditions Using Support Vector Machines on an Embedded Device

This article presents an innovative monitoring and prediction system for the conditions of Hass avocado crops, aiming to manage resources better and achieve greater productivity. It is designed to address the technological limitations of Peruvian agriculture. The project focused on integrating sensors that collect a large amount of environmental data, including temperature, humidity, pressure, light, soil moisture, and rain/snow, which are processed in real time on an ESP32 microcontroller. A dataset of 900 instances was used; the system was categorized into three groups: "Requires Attention," "Optimal," and "Critical Range." A historical and georeferenced database was used; the Support Vector Machine (SVM) model identified temporal and spatial patterns, providing key perspectives for optimal irrigation schedules and the long-term effects of weather on plant health. By having an SVM model with a linear kernel, the system achieved an accuracy of 97.78%. The robustness of the model allows for identifying crop conditions, highlighting the transformative potential of machine learning in Peruvian agriculture, in addition to contributing to smarter agricultural practices.