An electronic tongue which composed of selective electrodes is applied to mineral water recognition (optimization of sensor array). The task of the system is to distinguish among five brands of mineral water. For this purpose, various pattern recognition (PARC) procedures are employed: principle components analysis (PCA), independent component analysis, linked-like adaptive genetic algorithm (LAGA), et al. LAGA networks are proved to exhibit the best performance both in array optimization and mineral water recognition. Their further advantages, such as fast training and robustness, make them the suggested pattern classifiers for sensor array data.