A case study of event detection performance measure in WSNs using gini index

In event-triggered, outdoor applications of wireless sensor networks, the decision on the occurrence or otherwise of the event of interest is often influenced by the prevalence of errors and false alarms. Imprecise event detection decision (EDD) could prove costly for end users and processes that rely on the outcome of such EDD mechanisms. In this paper, we propose the application of Gini index as an EDD tool. We compare this method against the d' (d-prime) solution strategy in Signal Detection Theory (SDT). Results obtained illustrate the effectiveness and efficiency of our approach, serving as stimulus for further research on the subject. © 2013 by IFSA.