Crop Yield Prediction Using Machine Learning Algorithm

Crop Yield Prediction Using Machine Learning Algorithm

Prof.D. Manikannan, P. Soumya , B. Soumya , Gaurav Tripathi , N. Sohan Mani , L. Sharath , B. Soumya Sree

School of Engineering,  Malla Reddy University

ABSTRACT

Crop yield prediction is a critical aspect of agricultural planning and decision-making processes. Traditional methods heavily rely on historical data and expert knowledge, often resulting in limited accuracy and scalability. In recent years, machine learning (ML) techniques have shown promise in improving the accuracy and efficiency of crop yield prediction models. This study explores the fusion of diverse agricultural data sources—meteorological data, soil attributes, satellite imagery, and historical crop yields—employing various machine learning algorithms such as convolutional neural networks (CNNs), long short-term memory networks (LSTMs), and gradient boosting machines (GBMs) for predictive modeling. Emphasizing feature engineering and data fusion techniques, we assess the impact of temporal and spatial aggregation methods on model performance, while exploring attention mechanisms and ensemble learning to improve interpretability and predictive accuracy. Through comprehensive comparative analysis across different crop types and geographical regions, this research aims to address challenges of data heterogeneity and scalability, ultimately advancing the reliability and applicability of crop yield prediction models for sustainable agricultural practices. This research delves into the application of decision tree-based models, specifically focusing on their efficacy in predicting crop yields through the integration of diverse agricultural data sources. By utilizing decision trees, including CART (Classification and Regression Trees) and ensemble methods like Random Forests and Gradient Boosting, this study examines the fusion of meteorological data, soil characteristics, and historical crop yield records.