APPLICATION OF NEURAL NETWORKS TO ESTIMATE AVAILABLE IRON IN SOIL FOR MID AND NORTH IRAQI AREA
Mesopotamia Journal of Agriculture,
2011, Volume 39, Issue 4, Pages 69-74
AbstractThis a study was applied on forty 40 sites, 20 twenty of them in northern mosul soil and the other 20 in Baghdad region soils ,The different in the sites of sampling was taken into account for respect of available iron to plant and also different some chemical and physical characters for the soils. The object was to use technique of Neural Networks to find out mathematics model use to estimate the various variable iron for Mosul and Baghdad soils depending on some soil characteristics (Total-Fe, pH, OM, CaCO3, Sand, Silt, Clay) which were used as inputs for the assumed Neural Networks model to get deficient estimation for available iron in soil. The results of Neural Networks application was very good in terms of available iron Estimation depending of soil character signed above. Statistical analysis using linear Regression analysis between the suggested network output and the real data of available iron of soil samples indicate a very good relation ship. Coefficient of determination ( R2 = 0.95 ) , This indicate the efficient generalization of suggested artificial Neural Networks model in the soil of mid and northern of Iraq .
- Article View: 152
- PDF Download: 112