Training a NN usually requires a substantial amount of data, of which one part will be used for training and the rest will be used for testing the trained network. Having selected the significant process parameters and assigned them to network inputs, the next step is to acquire the required process parameter and response (output) data for training. It is important that the parameter data be correctly associated with the consequent process output data. Data collection is an important step to ensure the sufficiency and integrity of the data used to train the network, as the network performance can only be as good as the training data. It is not possible to say how many data items are appropriate, because this depends on the complexity of the process modelling problem. Generally the training data should be representative of the entire population of data items, unless there is good reason to resort to stratification or data blocking. The proportion of training to testing data varied considerably in the published research. For example Nascimento and Guardani (2000) vary the ratio from 1:1 to 3:1, Hsieh and Tong (2001) use 2:1 and Coit et al. (1998) use 4:1. Training and testing data can be acquired in a number of ways, as described in the following sections.