تخمین هدایت الکتریکی رودخانه ها با استفاده از شبکه عصبی موجک (مطالعه موردی: رودخانه کاکارضا)

     Electrical conductivity (EC) is an important factor in river engineering, especially studying of river water quality. In this study we studied and evaluated wavelet neural network to predict the electrical conductivity of the Kakareza river (in lorestan), and the results were compared with results of artificial neural network model. For this purpose, hydrogen carbonate, chloride, sulfate, calcium, magnesium, sodium and flow rate at monthly scale during the period (1969-2015) as input and output parameters were selected as electrical conductivity. The criteria of correlation coefficient, root mean square error and of Nash Sutcliff coefficient were used to evaluate and performance compare of models. The results showed that wavelet neural network model has the highest correlation coefficient (0.977), the lowest root mean square error (0.032 ds/m) and the highest standards Nash Sutcliffe (0.847) became a priority in the validation phase. The results indicate acceptable capability of wavelet neural network models to estimate the electrical conductivity of river water.