Shahid Beheshti UniversityEnvironmental Sciences1735-132413420151222Assessment among two data mining algorithms CART and CHAID in forecast air temperature of the synoptic station of ArakAssessment among two data mining algorithms CART and CHAID in forecast air temperature of the synoptic station of Arak535897635FAMehdi PanahiDepartment of Water Engineering, Faculty of Agriculture, University of Zanjan, ZanjanSeyyed Hassan MirhashemiFaculty of Water and Soil, University of Zabol, ZabolJournal Article20160227Due to recent droughts and global warming, in the modeling and prediction of climatic parameters it seems inevitable that the approach of using data mining algorithms to predict climatic elements has been widely used .Therefore, in this study the use of the CARD and CHAID data mining algorithms to predict the air temperature at Arak synoptic stations was evaluated. The data provided are the average monthly data from the Arak weather station, including: “sunshine hours”, “dew point”, “relative humidity”, “average wind speed” and “saturation vapor pressure deficit” over a forty-six-year period from 1960 to 2005. The output variable used was “average temperature months later” on a monthly basis. After introducing the weather data as a monthly mean to the algorithm as an input variable of air temperature months later. Then to the CART and CHAID algorithms were evaluated using correlation coefficient (R2) and mean absolute error (MAE). According to the two statistical indices, the CHIAD Tree model has a better function by R2 =0.915 and MAE= 2.77 in forecasting the monthly average temperature for the months later.Due to recent droughts and global warming, in the modeling and prediction of climatic parameters it seems inevitable that the approach of using data mining algorithms to predict climatic elements has been widely used .Therefore, in this study the use of the CARD and CHAID data mining algorithms to predict the air temperature at Arak synoptic stations was evaluated. The data provided are the average monthly data from the Arak weather station, including: “sunshine hours”, “dew point”, “relative humidity”, “average wind speed” and “saturation vapor pressure deficit” over a forty-six-year period from 1960 to 2005. The output variable used was “average temperature months later” on a monthly basis. After introducing the weather data as a monthly mean to the algorithm as an input variable of air temperature months later. Then to the CART and CHAID algorithms were evaluated using correlation coefficient (R2) and mean absolute error (MAE). According to the two statistical indices, the CHIAD Tree model has a better function by R2 =0.915 and MAE= 2.77 in forecasting the monthly average temperature for the months later.https://envs.sbu.ac.ir/article_97635_5a4406592e3f3f378ff8e54aa9678145.pdf