Wireless Sensor Networks (WSN) consists of several tens to hundreds of nodes, interacting with each other. Thus, they have multiple communications between them, transferring and receiving several packets of data to each other. In order to reduce the overall traffic in the network and lessen the presence of redundant node data, this paper proposes an adaptive sensor ranking method by evaluating the task necessity, utility, and region coverage of a particular node in a given WSN. Logistic regression has been used to adaptively train the WSN to assign a status to node as on or off, thereby, decreasing the overall data transmission into the network, while still accounting for the entire range of the WSN. © 2020, Springer Nature Singapore Pte Ltd.