Archives

International Journal of Pure & Applied Bioscience (IJPAB)
Year : 2018, Volume : 6, Issue : 6
First page : (212) Last page : (220)
Article doi: : http://dx.doi.org/10.18782/2320-7051.7145

Comparison of Abilities of Different Activation Functions of Artificial Neural Network to Predict Crop Area and Crop Production

Raju Prasad Paswan¹, Shahin Ara Begum² and Borsha Neog³

¹Department of Agril Statistics, Assam Agricultural University, Jorhat, India, 785013
²Department of Computer Science, Assam University, Silchar, India, 788011
³Department of of Agril Statistics, Assam Agricultural University, Jorhat, India, 785013
*Corresponding Author E-mail: rppaswan@aau.ac.in
Received: 7.10.2018 | Revised: 20.11.2018 | Accepted: 28.11.2018  

 ABSTRACT

India is a agricultural based country. The accurate information on crop area and crop production is very important for the development of Agriculture sector. It helps the planners to take certain decisions for solving the agricultural problem. Different Artificial Neural Network (ANN) models have been found to be suitable to predict crop area and crop production more accurately than the traditional equations. The present work has been carried out to find the best activation function of the multilayer perceptron (MLP) neural network to predict crop area and crop production of North Bank Plain Zone of (NBPZ) of Assam. The MLP models have been tested with three different types of activation function viz. Log-sigmoid, Hyperbolic tangent and linear transfer function. The performance of the developed models have been evaluated using Root Mean Square Error and Correlation Coefficient. It is found that ANN model with log-sigmoid transfer function in the hidden as well as output layer provides more accurate results compared to other configuration with the dataset considered.

Key words:Artificial neural network; Multilayer perceptron; Activation function; Crop area, Crop Production RMSE, CC.

Full Text : PDF; Journal doi : http://dx.doi.org/10.18782

Cite this article: Paswan, R.P., Begum, S.A. and Neog, B., Comparison of Abilities of Different Activation Functions of Artificial Neural Network to Predict Crop Area and Crop Production, Int. J. Pure App. Biosci.6(6): 212-220 (2018). doi: http://dx.doi.org/10.18782/2320-7051.7145