Assessment of different indices depicting soil texture for predicting chisel plow draft using neural networks [electronic resource].

Includes references.

The aim of the present study is to assessment of different indices depicting soil texture for predicting chisel plow draft using neural networks. So, six neural network models with different inputs and one output were trained using a backpropagation learning algorithm. The soil texture Indices were formed by different combinations of soil fractions. Available data In literature, directly related to our subject, were collected. These data were observations of field experiments. The input parameters were soil fractions In different forms (soil texture index), plowing depth. rated plow width, forward speed, initial soil moisture content, Initial loll bulk density and rated tractor power. The results showed that the neural network model with any soil texture index represented by soil fractions could predict chisel plow draft with reasonable accuracy. Correlation coefficients values between actual and predicted draft were higher than 0.80 for all neural network models However, values of mean absolute percentage error were 11.027 ole and 11.887 ole during training and testing the developed neural network model which used soil fractions values to represent soil texture as separated inputs, respectively.

Summary in Arabic.

1