Vol. 13 No. 1
Comparative study of feed-forward neuro-computing with multiple linear regression model for predicting mulberry leaf yield
Author(s): M. MISHRA, S. BHAVYASHREE AND G. C. GIRISHA
Abstract: In this study, the data of 16 mulberry genotypes for a year was collected from Department of Sericulture, UAS Bangalore. The results from multi-regression analysis using stepwise method for leaf yield showed that 3 variables (fresh leaf weight, number of leaves per plant and moisture content) having significant impacts. The main objective of this work is to compare the accuracy of Artificial Neural Network (ANN) and Multiple Linear Regression (MLR) models for prediction of leaf yield. We have compared MLR with ANN of two layer feed-forward network with sigmoid hidden neurons (10) using Levenberg-Marquardt (LM) algorithm. The performance of ANN was found to be better than the MLR model for leaf yield prediction with maximum value of R-square and minimum value of root mean square error.
Views: 790 | Downloads: 14