Key Nodes in Wireless Sensor Networks: Centrality Measures and Principal Component Analysis in Watts-Strogatz, Random Walk, and Barabási-Albert Models

Wireless Sensor Networks (WSNs) have facilitated the advancement of communication systems. WSNs are deployed in sectors where metrics and performance factors are calculated mainly based on particular nodes, the importance of which should be emphasized. Centrality measures are enumerated to include Degree, Betweenness, Closeness, Eigenvector, Katz and Subgraph Centralities, which directly assess the importance of individual network nodes. The research explicitly considers the generation of WSNs with 150 nodes using the Watts-Strogatz, Random Walk and Barabasi-Albert models, paying attention to essential nodes that affect the network level and its features. Using Pearson correlation analysis, Kendall rank correlation analysis and Spearman rank correlation analysis determine how centrality measures correlate in each model. Principal Component Analysis (PCA) determines how many nodes need to be used to determine the centrality data to contain the most variance across the different models. The findings in this study also underscore the importance of the centrality measures in explaining the network's topology and make it clear why some nodes are more important than others; the information provided may assist in creating more efficient WSN structures.