Check the Phenomenon of Dust in Khuzestan Province Using Decision Models CHAID and CRT

Aim of this study was to investigate the phenomenon of dust in the Khuzestan province and the correlation between these phenomena with elements of precipitation, speed wind and temperature is between 2010-1990. Statistical method library and using monthly data, eight selected stations (Masjed Soleiman, Ahvaz, Ramhormoz, Behbahan, Dezful, Aghajari, Abadan and Omidiyeh) of Khuzestan province using the softwares Minitab 17, spss 19 and Microsoft Excel 2013. For this purpose, using test homogeneity was examined data. From decision models was used to test, including automatic discovery classification method CHAID and regression trees CRT and test multiple interaction linear regression. Test Anderson-Darling and Kolmogorov-Smirnov showed that dust phenomenon data in Khuzestan province is normal. Using two methods CHAID and CRT was determined that wind is the best predictor for classifying dust in Khuzestan. Wind floors more than 2,100 in the CRT method and floor winds more than 6,500 in the CHAID method as a predictor of significant effect on the classification dust synoptic stations. Correlation between temperatur with dust is 0.425 percent, dust with wind speed 0.452 percent and between dust with rainfall -0.311 percent. With increase precipitation in the Khuzestan province reduced the number of days of dust but the temperature and wind speed are correlated to some extent with dust province and variables such as wind speed, rainfall and temperature have significant impact in the dust Khuzestan province.