Abstract
Precisely estimating crop yields is a critical aspect of agricultural planning, resource allocation, and food security. Satellite data integrated with machine learning algorithms have recently become a potential solution for predicting crop yield at local and global levels. The present study provides detailed investigation of satellite-based crop yield prediction using machine-learning algorithms. The proposed methodology integrates satellite imagery data with precipitation data. We use machine learning algorithms for predictive modelling, random forests, support vector machines, decision trees, linear regression, and k-nearest neighbour. Extensive investigations are conducted to examine the effectiveness of the proposed method. The study employs multi-year satellite imagery and corresponding crop yield data from various agricultural regions to develop predictive models. The models are trained and tested while considering temporal and spatial variations. Model accuracy and reliability are evaluated through performance metrics, including mean absolute error and root mean square error. The study’s findings indicate that using machine learning algorithms for satellite-based crop yield prediction yields a significant level of accuracy compared with standard techniques. According to the research conducted, it has been found that among all the methods that were implemented, the support vector machine method has shown better performance. Integrating satellite-based techniques and machine learning algorithms presents a viable and scalable approach to predicting crop yields.
Keywords: Crop Yield Prediction, Machine Learning, Remote Sensing, Satellite Imagery, Support Vector Machine.