Artificial Neural Network and Efficiency Estimation in Rice Yield
|Sanjib Kumar Hota
Lecturer, Madhusudan Institute of Cooperative Management, Bhubaneswar, Orissa, India
|Related article at Pubmed, Scholar Google|
The analysis of the effect of methodological difference in estimating the technical efficiency of production function of various agricultural farms drawn from different agro-climatic zones of odisha is the main objective of the study. The Data envelopment Analysis (DEA) and Artificial neural networks (ANN), viz. Multilayer Perceptron (MLP) and Radial Basis function (RBF) networks have been used in the study. The statistical software like Banixia frontier Analyst-7 (for DEA), MATLAB-ANN tool box (for MLP) and Neuro Solutions-6.0 have been used for estimating the technical efficiency through these models. The sensitivity analysis based on the optimum result of the network worked out by RBF has also been tested. The paper has illustrated the optimization problem by considering rice(paddy) crop yield data as the output (Y) and total Cost of Bullock/ Machine labor-per acre (X1), cost of human labor per acre (X2), Cost of seeds per acre (X3), Fertilizer cost per acre (X4), total-Irrigation Cost (X5), Cost of pesticide (X6) and Credit per acre (X7) and Gross cropped area under rice (X8) as inputs. It has been observed that the neural network-based estimation of technical efficiency may lead to a significant result; with radial basis function networks (RBFN) outperforming the other estimation techniques considered for the study. It is hoped that, in future, research workers would start applying these advanced ANN models to optimization problems relating to agricultural productivity.