PRECISION AGRICULTURE WITH SUPPORT VECTOR REGRESSION: HARNESSING AI ALGORITHMS IN MODERN SOFTWARE ENGINEERING

Abstract

This study explores the use of Support Vector Regression (SVR) in forecasting wheat yields within the scope of precision agriculture in Uzbekistan. In light of increasing climate variability and its effects on crop production, there is a growing need for machine learning models that can uncover non-linear relationships between environmental factors and agricultural outputs. The proposed methodology integrates SVR into a modular software pipeline using data collected from different agro-ecological zones of Uzbekistan between 2014 and 2030. Performance was assessed using RMSE, MAE, and R² metrics, with SVR achieving the highest accuracy (R² = 0.91) compared to Linear Regression, Decision Tree, and Random Forest. The results highlight SVR’s capability to generalize well under both normal and extreme conditions, offering valuable insights for sustainable agricultural planning in data-scarce environments. This work supports the development of AI-powered forecasting systems tailored to Uzbekistan’s agricultural needs.