Data Mining Process for Integrated Evaporation Model


JOURNAL OF IRRIGATION AND DRAINAGE ENGINEERING, cilt.135, ss.39-43, 2009 (SCI İndekslerine Giren Dergi) identifier identifier


The purpose of this study is to develop an integrating evaporation estimation model using a data mining process for the Lakes District in the southern part of Turkey. Lakes Egirdir, Kovada, and Karacaoren Dam are located in the Lakes District. The proposed data mining process is applied on these lakes for evaporation estimation. The daily pan evaporation data used in the data mining process are taken from State Hydraulic Works in southern Turkey. These data cover an 8-year period between 1998 and 2005 inclusively for daily pan evaporation of Lakes Egirdir, Kovada, and Karacaoren Dam. It is known that a developed integrated daily pan evaporation model is necessary for these lakes, which are so important to the Lakes District. Therefore, a data mining model having two inputs and one output is developed. Input parameters used in the developed models for Lakes Egirdir, Kovada, and Karacaoren Dam were daily pan evaporation values of Lakes Kovada + Karacaoren Dam, Lakes Egirdir + Karacaoren Dam, and Lakes Egirdir Kovada, respectively. As a result, in comparing the developed models with measured daily pan evaporation values, the REP tree model has better agreement with measured daily pan evaporation than other models. The results show the developed model was more accurate.