Predictive modeling of mineral content in rice grains: Insights from support vector and adaboost regression

Abstract

Rice is a vital staple food for over half of the global population, particularly in developing countries, serving as a primary source of calories and essential nutrients. Among the critical micronutrients found in rice, iron (Fe) and zinc (Zn) are crucial for human health, playing significant roles in preventing malnutrition and supporting various physiological functions. This study explores the relationship between physiological traits of rice plants such as plant height, tiller count, panicle number, fresh weight, panicle weight, biomass, and grain weight and the Fe and Zn content in rice grains. The models were developed using advanced machine learning techniques, specifically Support Vector Regression (SVR) and AdaBoost Regression, to predict iron (Fe) and zinc (Zn) content in rice grains. These predictions were based on physiological parameters collected from the crops. The results indicate that the SVR model is more effective for predicting Fe content, while the AdaBoost model performs better for Zn content.