Crop Yield Prediction in IoT: A Hybrid Feature Selection Approach using Machine Learning Models

The Internet of Things (IoT) is a pivotal technology applied across various domains such as smart homes, innovative healthcare, and smart farming. Despite its potential, IoT-based devices face major challenges, especially in combating crop diseases, addressing unpredictable weather patterns, and predicting crop yields. Traditional approaches predict crop yields often struggle with accuracy and reliability under environmental conditions. This study aims to increase predictive accuracy in an IoT environment by developing a hybrid feature selection method based on the Wrapper method. Therefore, we use advanced Machine Learning (ML) techniques like Adaptive Boosting (AdaBoost) and K-Nearest Neighbors (KNN) for crop yield prediction. The proposed system was evaluated using multiple performance metrics, demonstrating that the hybrid method improved classification accuracy by approximately $4 \%$ more. Evaluation of KNN and AdaBoost algorithms ensures AdaBoost’s superior accuracy of $\mathbf{9 4 \%}$ and reliability of $\mathbf{9 5 \%}$. Simulation results highlight the efficiency of our approach in comparison to current practices, promising significant improvements in farming decision-making and sustainability measures.