Cloud Computing-Enabled Hybrid Machine Learning Models for Crop Yield Prediction

The combination of the cloud computing and the machine learning can be regarded as the prospective solution of the issues related to modern agriculture, where predicting the crop yields is more accurate and scalable. In this work, we introduce an integrated machine learning system that uses cloud computing to support data-driven, near-real-time crop yield prediction. To enable the algorithm to learn in addition to time-based trends as well as nonlinear relationship(s) in agricultural datasets, we will then propose a model that combines the ensemble learning techniques of random forest and gradient boosting with the deep learning architecture of long short-term memory (LSTM) networks. With the scalability of processing, cloud infrastructure offers an easy access to several sources of data like satellite images, weather, soil, and yield history among others. In experiments, Model is tested using large scale agricultural data and model performance is measured using as RMSE, MAE, R <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> score. Just like all other collaborative filtering (CF) algorithms, they fail to work as sole models and produce low prediction accuracy and computational performance. The findings indicate that cloud enabled hybrid models suggested in this paper offer good support to precision farming and data driven decision making in agriculture which will contribute in the rising food security and sustainability in the farming processes. Such modifications are to be implemented in the future work as embedding real time sensor data and adaptive learning processes to make more accurate predictions.