Entropy based correlation clustering for wireless sensor network in multi-correlated regional environment
Thanh Nga Nguyen Thi, Nguyen Kim Khanh, Son Ngo Hong · 2016
The existence of correlation characteristics bring significant potential advantages for the development of efficient routing protocols in wireless sensor network. However, there has not an effective approach to divide sensor nodes into correlation group in multiple correlation-areas. This research proposed a new simple method on evaluating joint entropy of multiple sensed data to identify multi-correlation areas. The definition of correlation region based on entropy theory is also proposed. Following, a correlation clustering scheme with less computation is developed using proposed method and definition. The results are validated with real data set.