پيش بيني ظرفيت تبادل كاتيوني خاك با استفاده از شبكه عصبي مصنوعي و رگرسيون چند متغيره در منطقه خضرآباد يزد

Prediction of Cation Exchange Capacity Using Artificial Neural Network and Multivariate Regression in Khezrabad Region

Design and analysis of land-use management scenarios requires detail soil data bank including CEC data. Although CEC can be measured directly, its measurement is especially difficult and expensive in the Aridisols of Iran because of the large amounts of calcium carbonate. Pedotransfer functions (PTFs) provide alternative methods by estimating CEC from more readily available soil data. Soil samples were taken from at 12 pedons in Khesrabad, Yazd Province. Measured soil variables included texture (determined by Bouyoucos hydrometer method), organic carbon (determined using Walkely and Black rapid titration) and CEC (determined using Bower method). Then, we applied the artificial neural network (ANN), multivariate regression (MR) and several published PTFs to predict CEC, using easily measurable characteristics of clay, sand, silt and organic carbon. The results showed that ANN method gave the best result followed by MR method and finally the PTFs. Regarding the inputs and coefficients of PTFs, other regression based models had different performance. Among these models, none of them had absolute performance. In conclusion, the result of this study showed that training is very important in increasing the model accuracy of one region.