Forecasting Of Monthly Precipitation Using M5 Model Tree And Classic Statistical Methods (Case Study: Synoptic Oroumieh Station)

Document Type : Technical Note (5 pages)

Author

Lecturer, Tabriz Azad University, Tabriz, Iran.

Abstract

This study carried out to estimate monthly rainfall data of Oroumieh that are assumed to be lost from 2006 to 2007 , by classic statistical methods and M5 model tree using the software Weka using Mahabad, Khoy, Salmas, Makoo and Tekab stations . Among the studied stations, Mahabad station (R = 0.90) had the highest correlation with Oroumieh station. 26 scenarios of nearby stations have been introduced to Weka software in estimating monthly precipitation of Oroumieh station that among scenarios, the scenario which was defined as the simplest and most accurate scenario, included three Mahabad, Makoo and Tekab stations with values of (MAE = 7.19, R = 0.90, RMSE = 9.64) because of the lower input parameters to the model. Among the classical methods, the single best estimator (SIB) method has been selected as the best method with the highest correlation coefficient and the lowest error (R = 0.90 , RMSE = 10.51 ,MAE = 7.07). M5 model tree had the best performance in estimating quantities of data (R = 0.91 ,RMSE = 9.94 , MAE = 7.29) and is considered as an alternative and applied method in the calculation of monthly precipitation data due to simple linear and understandable relationships .

Keywords

Main Subjects

References

Berry M, Linoff G (1997) Data mining techniques:for marketing, sales, and customer support. New York John Wiley and Sons
Solomatine DP and Xue Y (2004) Pr M5 model trees and neural networks: application to flood forecasting in the upper reach of the Huai river in China. Journal of Hydrologic Engineering, 9(6), Doi.org/10. 1061/(ASCE)1084-0699(2004)9:6(491)
Ditthakit P, Chinnarasri C (2012) Estimation of pan coefficient using M5 model tree. American Journal of Environmental Sciences 8(2):95-103
Emamifar S, Rahimikhoob A, Noroozi A A (2013) Daily mean air temperature estimation from MODIS land surface temperature products based on M5 model tree. International Journal of Climatology Int. J. Climatol. 33:3174–3181
Etemad-shahidi A, Mahjoobi J (2009) Comparis on between M5  model tree and neural networks for prediction of significant wave height in lake Superior. Ocean Engineering 36:1175–1181
Quinlan JR (1992) Learning with continuous classes. Proceedings of Australian Joint Conference on Artificial Intelligence. World Scientific Press: Singapore, 343–348
Rahimikhoob A, Asadi M, Mashal M (2013) A comparison between conventional and M5 model tree methods for converting pan evaporation to reference evapotranspiration for semi-arid  region. Water Resour Manage 27:4815–4826
Sattari M , Nahrein F, Azimi V (2013) Forecasting of daily evapotranspiration using artificial neural network model and the  model tree M5 case study: Bonab station. Iranian Journal of lrrigation and Drainage 1(7):104-113 (In Persian)
Iran-Water Resources Research
Volume 13, Issue 4
January 2018
Pages 179-183

Files

History

  • Receive Date: 22 December 2016
  • Revise Date: 21 April 2017
  • Accept Date: 26 April 2017

Share

How to cite

Statistics