AN EXEMPLAR-BASED LEARNING-MODEL FOR HYDROSYSTEMS PREDICTION AND CATEGORIZATION

The main purpose of this paper is to represent a new Exemplar-Based Le arning model and to apply this model for river flow estimation. The ce ntral idea of the model is based on a theory of learning from examples . This idea is similar to human intelligence: when people encounter ne w situations, they often explain them by remembering old experiences a nd adapting them to fit. To explore the stability and efficiency of th e model performance, a simple mathematical function is simulated by th e model. The model is then applied to extend the annual stream flow re cords according to the nearby stream flow stations and to classify the monthly flow by using the monthly rainfall and runoff information in the previous months. The results show that the model has better perfor mance than the traditional methods and the results demonstrate the pow er and efficiency of the model for the hydrological data analysis.