Node Clustering Based on Feature Correlation and Maximum Entropy for WSN
Min Woo Kim, Kyung Tae Kim, Hee Yong Youn · 2019
Recently, wireless sensor network (WSN) has been drawing a great deal of attention both in academia and industry. Numerous schemes have been developed to maximize the performance and reliability of WSN, and node clustering is commonly employed for efficient management of the sensor nodes. In this paper a novel node clustering scheme is proposed which is based on the correlation between the features collected from the nodes, while the features are weighted using the maximum entropy model. It allows efficient measurement of the similarity between the features, and thus proper node clustering is achieved. Extensive computer simulation demonstrates that the proposed scheme significantly outperforms the existing representative schemes in terms of Adjusted Rand Index, Fowlkes-Mallows Index, and relative effectiveness.