An Enhanced Machine learning data aggregation model for Efficient data processing in wireless sensor networks

Abstract

Wireless Sensor Networks has more resource constraint on executing efficient data transmission, and it needs a robust routing protocol. Data processing efficiency in WSN depends on scalability, energy consumption, and QoS optimization with overhead. Achieve improved WSN proposing Machine Learning Data Aggregation Model (EML-DA). The proposed EML-DA architecture will focus on hybrid CH selection and robust data aggregation. Hybrid Cluster formation and CH selection work based on Artificial Neural Networks and robust data aggregation work based on Independent Component Analysis (ICA) ML technique. Considering the parameters as bandwidth allocated, the distance between Sensor nodes, and Base Station (BS), residual energy ANN architecture solves the problem by choosing the best CH of each cluster. Independent Component Analysis (ICA) ML technique performs efficient data aggregation at the CH node of each cluster to reduce energy consumption and minimal overhead; ICA is computationally efficient and uses differential entropy to minimize redundant data. EML-DA outperforms current ML-based clustering and data aggregation algorithms, affording to the outcomes of the experiments.