Evaluation and Performance of Support Vector Machine Model in Estimation of Suspended Sediment

Sediment transport has been constantly influenced the river and civil structures and the lack of information about its exact amount causes high damages. Achieving to proper procedure is important to estimate the sediment load in rivers. This study used the support vector machine model to estimate the sediments of the Kakareza river placing on Lorestan Province and their results were compared with results obtained by gene expression programming. Parameters including river discharge, rate of dissolved solids and precipitation for time period (1993-2013) were monthly selected. Criteria including correlation coefficient, root mean square error and mean absolute error were used to evaluate and also compare the performance of models. The achieved results showed that combinational patterns using two intelligence models could be investigated and acceptable results were presented for sediment rats. With regards to accuracy, the support vector machine model showed the highest correlation coefficient (0.867), minimum root mean square error (0.024 ton/day) and the mean absolute error (0.017 ton/day) which was initiated at verification stage. Finally, the results showed that the support vector machine has been shown great capability to estimate the minimum and maximum values for sediment discharge.