A neural networks application in ergonomics

In this paper is presented a feed forward neural networks application in ergonomics. The neural network is used to quantify from the ergonomics point of view, a working place. Based on the combination of six input measurable parameters, the working place is characterized with one from the following three categories: “good”, “medium” or “poor”. The six input parameters that are taken into account by our application are: temperature, humidity, noise, luminosity, weight and frequency. The Java application crates the feed forward neural network, trains it using a set of training patterns and then tests it using real data. The experiment and the Java application were made at University of Pitesti.