|
PRINT ISSN : 2319-7692
Online ISSN : 2319-7706 Issues : 12 per year Publisher : Excellent Publishers Email : editorijcmas@gmail.com / submit@ijcmas.com Editor-in-chief: Dr.M.Prakash Index Copernicus ICV 2018: 95.39 NAAS RATING 2020: 5.38 |
Accurate forecasting of prices of bivoltine mulberry cocoons is essential for planning and policy purposes. A study had been taken up to develop an appropriate model for forecasting the daily prices of biovoltine reeling mulberry cocoons by using the data collected for the period from 1st April 2010 to 31st of August 2019 from Government Cocoon Market (GCM), Ramanagaram, Karnataka. Generalized Auto Regressive Conditional Heteroscedastic (GARCH) and Artificial Neural Network (ANN) models were used to analyze the past behavior of time series data in order to make inferences about its future behavior for prices of bivoltine cocoons. A suitable model of GARCH process was identified based on the results of Akakie Information Criterion (AIC) and Bayesian Information Criterion (BIC).GARCH (3, 2) model exhibited lesser AIC and Schwarz criterion compared to other process. The estimates of GARCH (3, 2) model showed that the coefficients of mean and variance equation were statistically significant at both 1% and 5% level of significance. A multilayered perceptron neural network was fitted, which had an input layer with 15 input nodes, 1 hidden layer with 10 hidden nodes and an output layer with one output node. The forecasting performance was assessed for both the models by using Mean Square Error (MSE), Root Mean Square Error (RMSE) and Mean Absolute Percentage Error (MAPE) and found that the fitted GARCH (3, 2) model was found to be better than ANN model.