Neural Approach Applying for Soil Classification System

Abstract

The main aim of this paper is to develop a system for classifying the soil class of Myanmar by the help of Error Back-propagation algorithm (EBP) which is the most widely used algorithm among Artificial Neural Network (ANN ) technique. This system, “Soil Classifier” includes two parts in general, training and testing. During the training phase, the Soil Classifier accepts nine inputs. These inputs are soil type, land use type, land form, soil depth, soil texture, soil PH and the percentages of each of three types of plant nutrients. The plant nutrients are Nitrogen ( N ), Phosphorus ( P ) and Potassium ( K ). After accepting the nine inputs, the Soil Classifier will generate one of three types of outputs whether the soil is good class, fair class or poor class. The Soil Classifier uses Multi Layer Feed-forward Neural Network.