Precision agriculture using voting ensemble method based on weather and soil conditions

Precision Agriculture, essential for optimizing crop yield and minimizing waste in the face of global population growth, relies on technology and data-driven approaches. Applying a machine learning model, that incorporates specific weather and soil data, stands out as a pivotal strategy in enhancing the efficiency of precision agriculture. Yet, constructing a singular optimal model poses challenges, necessitating a nuanced approach to achieve precise farming objectives. To address this, we propose a precise predictive model using hard voting ensemble method, aiding farmers in informed decision-making regarding crop selection and management, cost reduction, and profit maximization. Our ensemble integrates K-Nearest Neighbors, Random Forest, and Gradient Boosting models, leveraging environmental factors such as rainfall, humidity, and phosphorus levels to determine the most suitable crop type. The model demonstrated exceptional accuracy, highlighting its efficacy in agricultural decision-making. Notably, the model exhibited a remarkable overall accuracy of 99.26%, precision of 99.12%, and f-scores of 99.03%. These robust results underscore the model's reliability, affirming its potential to significantly enhance crop selection decisions and optimize yields in precision agriculture.