A Hybrid Spatio-Temporal and Machine Learning-Based Data Aggregation Model for Redundancy Reduction in Wireless Sensor Networks

Wireless Sensor Networks (WSNs) suffer from excessive data redundancy due to dense deployment and overlapping sensing regions. Traditional data aggregation techniques reduce redundancy but often fail to balance energy efficiency, accuracy, and adaptability. This paper proposes a Hybrid Spatio-Temporal Machine Learning-based Data Aggregation (HST-MLDA) model that integrates clustering, correlation analysis, and lightweight machine learning techniques to eliminate redundant data effectively. The proposed model combines temporal filtering at the node level and spatial correlation at the cluster level, followed by classification-based redundancy elimination. Simulation-based analysis demonstrates improved network lifetime, reduced energy consumption, and enhanced data accuracy compared to traditional aggregation techniques.