Comparison of Artificial Neural Network and Regression Models for Estimating Dry Weight and P Uptake of Corn

In this study، a comparison between the artificial neural network (ANN) and linear regression models for estimating the dry weight of corn and its P uptake، based on the extracted P from soil by different extractants was done. For this purpose، 25 surface soil composite samples (0-30 cm) were collected from different points of East Azerbaijan province، and then corn plants (single cross 704) were cultivated in these soils under the greenhouse condition with three replications. After 60 days، the plants were harvested and the shoot dry weight and its P concentration were measured. The results showed that the coefficient of determination (r2) values between the extracted soil P، obtained by Colwel and Olsen''s tests and corn shoot dry weight were 0. 49 and 0. 44، respectivly. The results of ANN showed that the Olsen''s test for estimating the corn shoot dry weight and distiled water for estimating the shoot P concentration were supreoir. For prediction of the important indices of the corn shoot dry weigh and P uptake، based on P concentration measured using different extractions، higher values for coefficient of determination were obtained by applying some conventional methods of ANN with respect to those obtained by applying linear regression methods، so it was concluded that ANN could be used in soil P testing.