Abstract
This paper describes the reduction of the number of nodes in a WSN network using a neural network. The aim is to show that the neural network can predict the measured quantity at a given location of the wireless sensor network based on measurements at other points. After successfully setting the parameters of the neural network, the node can be removed. The theoretical part deals with wireless sensor networks and an introduction to feedforward neural networks. The experimental part describes the prediction of the temperature at a given point based on measurements at the three surrounding points of the WSN network.