Predicting the monthly flow of the Kaleybar Chay River based on M5 model tree

Abstract

Due to its importance in the designing of water projects, flood investigation and estimation has always been one of the key issues in the field of hydrology and researchers have tried to estimate the river flow more accurately by using different methods. In this regard and in the present study, the monthly discharge values of Kaleybar Chay River were predicted using machine learning method of M5 model tree. Based on this, time series data of discharge and precipitation in monthly delays are used as input parameters of the models and the results are evaluated using the statistical indicators of correlation coefficient, root mean square error and mean absolute error. Finally, the values obtained from the M5 models showed that among the studied models, M5(6) model with the root mean square error of 0.968 and the mean absolute error of 0.625, had the highest correlation with the observed values and recorded the most accurate results in this study. It was also found that most of the M5 models had a successful performance in estimating the monthly flow in the study area.