ML-Based Yield Prediction in Smart Agriculture Systems Using IoT

The Internet of Things (IoT) and machine learning (ML) are two of the most rapidly expanding academic areas. “Smart x” systems that utilize ML and IoT include smart houses, smart cars, smart campuses, smart cities, and early warning systems. The worldwide flow of information between organizations will change as a result of the use of such systems. This research provides a hybrid ML model that utilizes IoT to forecast crop yields. The three main parts of this research are preprocessing, classification, and Smart Campus (SC). Using the Variance Inflation Factor (VIF) technique and correlation-based sensitivity analysis (CBSC), we perform FS after cleaning the data. Last but not least, we propose a two-tiered machine learning architecture for IoT-enabled smart farming. The Adaptive k-Nearest Centroid Neighbor Classifier (aKNCN) model is presented in the first tier to classify soil samples into groups according to the input soil parameters and to calculate the soil quality. Using the Extreme Early Warning System (EWS), the crop yield is forecasted in the second tier. Utilising the Enchance Butterfly Optimisation algorithm (eBOA), the weights are modified in the optimised strategy to increase ELM performance accuracy while minimising error numbers. The application tool for testing the suggested system is Python. The suggested prediction model’s performance is assessed using the soil dataset. The performance evaluation considers several parameters, including accuracy, RMSE, R2, MSE, MedAE, MAE, MSLE, MAPE, and Explained Variance Score (EVS).