An Artificial Neural Network Approach to Forecasting

R. Ahluwalia and S. Chidambaram (USA)


Neural Networks, Back Propagation, Linear Regression


Back propagation is the most widely used procedure to implement artificial neural networks. This paper describes a five-step (data input, data scaling, forward pass, backward pass, and output) procedure to implement the back propagation algorithm. An illustrative example is used to compare predictions from an artificial neural network model with predictions from a linear regression model.

Important Links:

Go Back