摘要：The soft computing techniques have gained popularity for predictions. Stage discharge studies play a crucial role in planning, design or management of any hydraulic system. Over or under estimation of discharge value causes huge loss of investments, structures and lives. Two neural networks have been studied taking stage discharge data of an Indian river named Brahmani. Performance of each network has been summarized. Accuracy of each network model is based on the percentage of successful predictions on the test sets of each data set. Accuracy is measured via the holdout method as well as through cross validation. The present work suggests the suitability of a neural network as a tool for predicting discharge which will be useful in different field of science and engineering.
关键词：Artificial neural networks;Elman network;Cascade network;stage and discharge
