FORECASTING OF FUTURES TRADING VOLUME OF SELECTED AGRICULTURAL COMMODITIES USING NEURAL NETWORKS
KrishiKosh
View Archive InfoField | Value | |
Title |
FORECASTING OF FUTURES TRADING VOLUME OF SELECTED AGRICULTURAL COMMODITIES USING NEURAL NETWORKS
|
|
Creator |
NAVYA NANAIAH, B
|
|
Subject |
apples, markets, retail marketing, supply, sales, marketing, economics, productivity, area, storage
|
|
Description |
A study of futures trading of agricultural commodities and the volume of their trade was done by using forecasting methodologies. The identified forecasting methodologies for forecasting futures trading volume of selected agricultural commodities are Regression, ARIMA and Neural networks. Monthly transactions data for the period of 2003 - 2010 has been collected from National Commodity & Derivatives Exchange (NCDEX) website for some of the agricultural commodities like maize, mustard, soybean and pepper. Volume was taken as dependent variable while open interest, closing price and mean cash price were taken as independent variables for analysis. The data obtained for past seven years on volume of commodities selected is used to predict the future volumes for the first five months of 2011. Model validation was done using MAPE, RMSE and Theil’s U statistic .The results indicated that the neural networks outperforms multiple regression and ARIMA. This study also helps the policy makers to ascertain the future requirements. Most of the data agriculture will be having non linear relationship and hence ANN modeling is most suitable for prediction of agriculture system. |
|
Date |
2016-12-15T10:54:35Z
2016-12-15T10:54:35Z 2011-03-23 |
|
Type |
Thesis
|
|
Identifier |
TH-10126
http://krishikosh.egranth.ac.in/handle/1/90265 |
|
Language |
en
|
|
Format |
application/pdf
|
|
Publisher |
University of Agricultural Sciences GKVK, Bangalore
|
|