Abstract

Feed forward neural networks are typically used in pattern recognition applications. They are also used for the approximation of nonlinear functions. In this paper we use a feed forward neural network for the prediction of the monthly consumption of gas in a house. The neural network is trained using the data about the monthly consumption of gas in the previous three years. We use as N inputs in the network the consumption of N previous months, in order to calculate the output of the network, the consumption of gas for the next month. The simulations are done using Java language. In the last section of this paper are presented the experimental results, including the approximation error.