Aquifer parameters are the important factors for assessing groundwater potential in any area. Yet estimation of aquifer parameters is expensive and time-consuming. This study proposes an optimal and improved model to make a quantitative and qualitative correlation between pumping test data set and aquifer parameters by integration of artificial neural network training algorithms and the supervised committee machine concept. This supervised committee machine with training algorithms (SCMTA) combines Levenberg–Marquardt (LM), Bayesian regularization (BR), gradient descent (GD), onestep secant (OSS) and resilient back-propagation (RP) algorithms using a supervised combiner to estimate non-leaky confined aquifer parameters using pumping test data set. Each of these algorithms has a weight factor showing its contribution in overall prediction. The results reveal that RP, BR and LM have more contribution than OSS and GD. The developed SCMTA model trained with 2000 training sets of the Theis well function and tested with 800 sets of synthetic time-drawdown generated from the aquifers parameters. In situ observation data from the time-drawdown at station Katasbes in Shiraz plain, Southwest of Iran, are further adopted to test the applicability and reliability of the proposed method. The results of this study suggest that the SCMTA model performs better than the individual artificial neural networks differing in training algorithms, the simple averaging and weighted averaging committee machine methods and the type-curve matching technique. Additionally, results indicate that the SCMTA method corrects the concept of the superimposed plot by applying a supervised combiner to determine the optimal match point and estimate aquifer parameters more precisely. The proposed SCMTA method is recommended as an alternative to the type-curve graphical method and the existing ANN approaches.
