The aim of this project is to develop a machine learning-based system for predicting sugar yield from sugarcane. The dataset used contains various agronomic features such as sugarcane genotypes, crop stages, population count, and stalk weight, among others. By leveraging machine learning models, including Stacking Regressor, Voting Regressor, CNN, AdaBoost, and CatBoost Regressor, the system generates accurate predictions for sugar yield, represented by T_SPACRE, which indicates the pounds of sugar per acre. This prediction model offers insights into the factors influencing sugar production and helps optimize breeding programs for better yield outcomes. The system is built on a Flask-based web application, where users can input crop data to obtain predictions. The results from different machine learning models are compared to select the most effective approach for yield prediction. The project demonstrates the power of machine learning in agricultural research, providing an efficient and automated way to forecast sugarcane productivity.

Keywords: Machine Learning, Sugar Yield, Sugarcane, T_SPACRE, Stacking, Voting, CNN, AdaBoost, CatBoost, Flask.