Design of ANN Based Classifiers for Soil Fertility of Uttarakhand

Abstract

The present article is based on the application of artificial neural network (ANN) in pattern recognition and classification. The ANN belongs to the family of AI (Artificial Intelligence). In general, an ANN is composed of single or multiple hidden neuron layers between input and output neuron layers. The way a human brain is trained, the training of an ANN is also performed exactly in the same way by showing it some input patterns. There are a variety of training algorithms for an ANN. In the present research two different training algorithms are used viz. competitive learning, and the scaled conjugate gradient (SCG) algorithm. Classification models are designed in MATLAB using these two algorithms for carrying out soil classification into two classes viz. class 1 (More Fertile) and class 2 (Less Fertile) based on the input pH value and available potassium content. The classification accuracy of both models is analyzed and compared.