Author's reply to discussion of using artificial neural nets to identify the well-test interpretation model

The authors thank Yeung et al. for their discussion about their original paper. They agree with Yeung et al. that their proposed scaling method, when applied to patterns with distinct subparts such as the one shown, represents an improvement on the method they proposed. This is particularly true because Yeung et al.'s method eliminates the need to train the artificial neural networks (ANN's) on different sizes (scales) of the same pattern of a specific interpretation model. This paper presents the following comments for discussion and suggestions for further improvement.