Artificial Neural Network Based Modeling of Spatial Distribution of Phosphorus on the Tomato Area

Abstract

In this paper, an artificial neural network (ANN) based modeling has been presented. Simulation of spatial variability of available phosphorus levels for soils on the tomato crop area was done by using the model of ANN. The method is commonly successful in issues such as model selection and classification, function forecast, determination of optimum value and data classification. In this study, a program was developed in C++ programming language to analysis the data by means of ANN method. For this aim, topsoil (0-20 cm), subsoil (20-40 cm) and plant samples were taken from the tomato plots based on 20 meters period and the area was modeled with 5 meters period. The findings clearly showed that a great spatial variability occurred in available phosphorus for topsoil and subsoil. In order to observe the performance of ANN based phosphorus model, experimental data curves have been compared with the curves obtained after the ANN training. After comparing the experimental results with the recommended method, it is observed that it shows realistic results. The results obtained from the simulation show that the modeling which is formed for the available phosphorus change is applicable. Therefore the developed model could be an alternative method for the predictions of phosphorus in tomato crop areas since the created neural network model of phosphorus resembles the actual data.