The processing of data and signals provided by sensors aims at extracting relevant features which can be used to assess and diagnose the health state of the monitored targets. Nevertheless, wireless sensor networks (WSNs) present a number of shortcomings that have an impact on the quality of the gathered data at the sink level, leading to imprecise diagnostics of the observed targets. To improve data accuracy, two main critical and related issues, namely the energy consumption and coverage quality, need to be considered. In this paper, we present a distributed algorithm based on a theory of domination in graphs, and we study its impact on diagnostics by using six machine learning algorithms. First, we give the correctness proofs and next we assess its behaviour through simulations. Obtained results show that the proposed algorithm exhibits good performances despite its lower complexity.