Long term water demand forcasting : use of Monte Carlo cross validation for the best model selection

摘要: Selection and validation of any statistical models are very crucial in modelling forecasting problems. In multiple regression analysis long term water demand, various developed with a variety predictor variables. Moreover, can take different forms such as linear, semi-log log-log. this paper, an effective but simple procedure named Monte Carlo cross (MCCV) is applied compared to the most widely used leave-one-out (LOO) select best model forecast demand. Unlike LOO validation, MCCV leaves out major part sample during validation. Both methods also for estimating prediction ability selected on future samples. The advantage that it reduce risk over fitting by avoiding unnecessary large model. demand data set Blue Mountains, NSW Australia single dwelling residential sector. results show has appropriate It found that, assesses higher degree accuracy. Furthermore, provides less uncertainty when