A Hybrid SVM–ABC Model for Monthly Stream Flow Forecasting

Abstract

AbstractDeveloping reliable estimates of stream flow forecasting is crucial for water resources management and flood forecasting purposes. The study investigates accuracy of hybrid support vector machine (SVM) integrated with artificial bee colony optimisation (ABC) algorithm for monthly stream flow forecasting. ABC algorithm is applied for improving performance of SVM model by helping in selection of optimal SVM parameters. Monthly flow data from two stations (Salebhata and Sundargarh) on Mahanadi River basin, India, are used in the study. Results of hybrid model are compared with results of conventional SVM model, and quantitative statistical measures are used for validating both models. Obtained results prove that hybrid SVM–ABC model processes complex hydrological data better and have better generalisation capability with higher prediction accurateness.KeywordsABCMahanadi riverStream flowSVM