Data driven modelling and knowledge discovery in water resources engineering

Data driven approaches have the potential of becoming highly useful knowledge discovery tools when domain knowledge is incorporated into learning procedure. This research defines a Genetic Programming (GP) based conceptual modelling framework coded in an open source R environment to understand catchment scale hydrological processes. The state of the art applications of GP in hydrology involve the use of GP as a short-term prediction and forecast tool rather than as a modelling framework. In this study, GP simultaneously evolves suitable model structures and associated parameters in a readily interpretable form, that explain set of observations with the help of background knowledge. Thus evolved GP model configurations are found to be in good agreement with fieldwork evidence.