Evaluating Machine Learning Models for Wheat Yield Prediction in Amritsar District

Agriculture is pivotal for global food security and economic stability, yet optimizing yields remains a formidable challenge amidst variable climatic conditions. Accurate and timely crop yield prediction (CYP) is crucial for strategic decision-making, particularly in countries like India, where agriculture is a lifeline for a substantial segment of the population and the economy. The advent of modern technology, notably machine learning (ML), offers promising solutions to the complexities of CYP. ML techniques provide an advanced alternative to conventional statistical methods, revealing complex patterns within datasets, particularly in cases involving nonlinear relationships. This study harnesses meteorological data to forecast wheat yields in the Amritsar district of Punjab. Through the application and comparison of six ML models, we endeavor to ascertain the most efficacious method for CYP. Mean Absolute Error (MAE), Root Mean Square Error (RMSE), Coefficient of Determination (R 2), and Mean Average Percentage Error (MAPE) are employed for model validation. Our results indicate that ensemble ML models significantly surpass linear ML models, demonstrating the lowest error metrics. An exhaustive comparative analysis of all models is conducted, providing valuable insights to the readers. Consequently, this study underscores the potential of ML techniques to augment CYP accuracy. The integration of cutting-edge technologies into agricultural practices enables stakeholders to make enlightened decisions, thereby fostering sustainable agricultural production and bolstering food security in India and beyond.