%0 Journal Article
%T Modeling of yield and rating of land characteristics for corn based on artificial neural network and regression models in southern Iran
%J Desert
%I University of Tehran
%Z 2008-0875
%A Zeinadini Meymand, A.
%A Bagheri Bodaghabadi, M.
%A Moghimi, A.
%A Navidi, M.N.
%A Ebrahimi Meymand, F.
%A Amir pour, M.
%D 2018
%\ 06/20/2018
%V 23
%N 1
%P 85-95
%! Modeling of yield and rating of land characteristics for corn based on artificial neural network and regression models in southern Iran
%K Critical production
%K Crop requirements
%K land suitability
%K Corn
%K Threshold limit
%K Very hot region
%R 10.22059/jdesert.2018.66355
%X This study was conducted to rate the land characteristics of corn in hot areas based on artificial neural networks and regression models. For this purpose, 63 corn fields were selected in southern Iran. In each farm, a pedon was excavated, described and sampled. A questionnaire was completed for each farm. A stepwise regression model was used to study the relationship between land characteristics and corn yield. A characteristic-function curve was used to rate the land characteristics. Finally, crop requirements were prepared by artificial neural network and regression models and verified by comparing the actual and predicted performance levels. The results of regression analysis showed that soil salinity, exchangeable sodium percentage, sand, clay, phosphorous, gypsum and potassium recorded the highest effect on yield and according to the artificial neural network, the exchangeable sodium percentage, soil salinity, soil texture and cation exchange capacity are the most important. Based on regression and artificial neural network methods, the threshold limit and break even production for soil salinity were 4, 2.5, 12, and 10 dS m-1, respectively, but for exchangeable sodium percentage the values were 18, 14, 35, and 30, respectively. The coefficient of determination (R2) between the actual and predicted yield based on the regression model was 0.88, but it was 0.945 (training data) and 0.837 (testing data) for the artificial neural network. Also, the results of the verification of the prepared crop requirements tables showed that the correlation of determination between the land index and the yield in the regression method was 0.78 but it was 0.81 for the artificial neural network, these results are acceptable in both methods.
%U https://jdesert.ut.ac.ir/article_66355_d8a1f991ec5ad1bc76232cdec5635dc0.pdf