Using soft computing and machine learning algorithms to predict the discharge coefficient of curved labyrinth overflows

This research aims to estimate the overflow capacity of a curved labyrinth using different intelligent prediction models, namely adaptive neural-fuzzy inference system, support vector machine, M5 model tree, least-squares machine and machine–bat algorithm (LSSVM-BA). A total 355 empirical data for 6 congressional models were extracted from results laboratory study on models. The parameters upstream water head ratio, lateral wall angle curvature used discharge coefficient overflows. Based various statistical evaluation indicators, show that those input can be relied upon predict coefficient. Specifically, LSSVM-BA showed best accuracy during training test phases. Such low-cost may have remarkable practical implication as it could an economic alternative expensive solution, which is costly time-consuming.