h. shekofteh

Soil quality indicators are measurable characteristics of the soil affecting the soil capacity for crop production or environmental performance. Among these indicators, air capacity (AC) and relative field capacity (RFC) are believed to be the most important ones. To select the best combination that affects soil physical quality indicators (AC and RFC), we employed a hybrid algorithm: an ant colony organization (ACO) in combination with an artificial neural network (ANN). Multiple linear regression and support vector regression models were constructed for the comparison of performances. The results obtained from running ACO-ANN to select the best combination revealed that a combination with four input variables, including soil organic matter, clay, carbonate calcium equivalent, and bulk density, had the lowest error. The R2 values in the ACO-ANN model for the AC and RFC predictions were respectively 0.91 and 0.95 whereas they were 0.75 and 0.53 respectively in support vector regression model, and 0.54 and 0.53 in the multiple linear regression model. Since the results obtained from the ACO-ANN algorithm are acceptable, this algorithm could be applied to other locations of the world in order to tackle environmental problems.  The results form sensitivity analysis for the ANN model showed that carbonate calcium equivalent and clay content had the highest and the lowest effects on AC and RFC indicators, respectively.