AHSWN Home • Issue Contents • Forthcoming Papers

Gradient Boost-Modified Classifier with Particle Swarm Optimization and Stochastic Diffusion Search Method for Data Optimization in Wireless Sensor Networks
E Suganya and S Sountharrajan

Wireless Sensor Networks (WSNs) embodied voluminous economical sensor nodes over a wide range of zones for the collection of data. The nodes are preliminary components of the Internet of Things (IoT). In WSN, resource constraints occur in several ways on the nodes; a few of them are resources of energy, computing, storage, data aggregation, and traffic delay.To maintain a stable network for a longer duration of time, it is essential to have a sturdy routing protocol, which helps to achieve greater energy utilisation. Clustering algorithms play a vital role in maintaining the energy efficiency of nodes. The performance of the clustering algorithm will be according to the locus of the energy centroid and the residual energy of nodes. One such prominent routing protocols is the Energy-Efficient Low-Energy Adaptive Clustering Hierarchy protocol (EE-LEACH). A substantial count of protocols have been designed recently, to enhance the performance of LEACH. This is being achieved either by the detection of multi-hop paths from the Cluster Heads (CHs) to the Base Station (BS) or by the reduction of the energy of the CH. The proposed method is designed to use the Gradient Boosting (GB) classifier, the modifying GB classifier with Particle Swarm Optimization (PSO), the Genetic Algorithm (GA), as well as the Stochastic Diffusion Search (SDS) for classification procedures. The current research proposes an ensemble algorithm, PSO with GB, GA with GB, SDS with GB, and SDS-GA with GB, as a classifier for the prevention of repetitive classifiers. The results of the experiments demonstrated the proposed SDS-GA-GB efficiency for data optimization and the proposed EE-LEACH SDS-GA for network performance improvement.

Keywords: Internet of Things (IoT), Wireless Sensor Network (WSN), Big Data, Routing, Particle Swarm Optimization (PSO), Genetic Algorithm (GA), Stochastic Diffusion Search (SDS), and Gradient Boosting (GB) Classifier

Full Text (IP)
DOI: 10.32908/ahswn.v54.9811