Ordinal-set pair analysis prediction model and application in Liao River basin

Abstract

Based on the combination of the ordinal theory and the set pair analysis theory, a novel method called ordinal set pair analysis model (O-SPAM) is proposed to simplify the process of runoff prediction and improve the prediction precision. The annual runoff data of fifty-two years from 1956 to 2007 of a certain reservoir in Liao River basin are used to prove the prediction effect of O-SPAM. The results indicate that seven relative errors which account for 70% of the total prediction number are less than 5%; one which accounts for 10% of the total number is between 5 and 10%; while the others which account for 20% of the total number are between 10 and 15%. All the relative errors are less than 20%, which meets the required precision of the annual runoff prediction in "Standard for hydrological information and hydrological forecasting", indicating that the model can predict the runoff accurately. The study demonstrates that O-SPAM is of clear principle, simple structure, easy calculation, and good prediction precision, and the model is applied to the annual runoff prediction. © 2012 Springer Science+Business Media B.V.