CorrACC-guided feature selection and metaheuristic hyperparameter tuning for deep learning in IoT Environmental monitoring
Department of Computer Engineering, Faculty of Electrical and Computer Engineering, Hakim Sabzevari University, Sabzevar, Iran.
Research Article
Open Access Research Journal of Engineering and Technology, 2026, 10(02), 037-057.
Article DOI: 10.53022/oarjet.2026.10.2.0027
Publication history:
Received on 25 March 2026; revised on 24 April 2026; accepted on 27 April 2026
Abstract:
The Internet of Things (IoT) has significantly advanced environmental monitoring by enabling continuous data acquisition from distributed sensor networks. However, challenges such as high-dimensional feature spaces, redundant measurements, and limited computational resources hinder the efficiency and reliability of predictive models. This study proposes an integrated learning framework that combines correlation-aware feature selection (CorrACC), metaheuristic hyperparameter optimization, and deep learning for accurate and efficient environmental prediction. The framework systematically evaluates both classical machine learning models (KNN, Decision Tree, Random Forest, and Lasso) and deep learning architectures (CNN, RNN, LSTM, and Bi-LSTM) under a unified experimental setup. To enhance model performance, Genetic Algorithm (GA), Particle Swarm Optimization (PSO), and Differential Evolution (DE) are employed for automated hyperparameter tuning. Experimental results demonstrate that deep learning models consistently outperform traditional approaches, with Bi-LSTM achieving the highest predictive accuracy across multiple evaluation metrics. Furthermore, the integration of metaheuristic optimization improves convergence stability and reduces training time, with DE showing the most consistent performance gains. The proposed framework effectively reduces feature redundancy while preserving critical information, leading to improved scalability and computational efficiency. These findings highlight the potential of hybrid heuristic–deep learning approaches for robust, real-time IoT-based environmental monitoring in smart and sustainable systems.
Keywords:
Internet Of Things (IoT); Environmental Monitoring; Feature Selection; Metaheuristic Optimization; Deep Learning
Full text article in PDF:
Copyright information:
Copyright © 2026 Author(s) retain the copyright of this article. This article is published under the terms of the Creative Commons Attribution Liscense 4.0