Data Pre-processing Methods and their impact on ANN

In analysis with ANN many times noisy data which is non-stationary comes into consideration for time series prediction. This book goes into different data pre-processing techniques and how these actually help in increasing the prediction accuracy of Artificial Neural Networks. The choice of the correct configuration of an ANN is a non trivial task and the results obtained are verified by applying various criteria and interchanging of experimental and test sets and randomizing so that these are applicable in extreme circumstances.