RAMESH DASYAM1*, BANJUL BHATTACHARYYA2, P. MISHRA3
1Department of Agricultural Statistics, Bidhan Chandra Krishi Viswavidyalaya, Mohanpur, Nadia, 741252, West Bengal, India
2Department of Agricultural Statistics, Bidhan Chandra Krishi Viswavidyalaya, Mohanpur, Nadia, 741252, West Bengal, India
3Department of Agricultural Statistics, Bidhan Chandra Krishi Viswavidyalaya, Mohanpur, Nadia, 741252, West Bengal, India
* Corresponding Author : pradeepjnkvv@gmail.com
Received : 05-10-2016 Accepted : 08-10-2016 Published : 01-11-2016
Volume : 8 Issue : 53 Pages : 2782 - 2787
Int J Agr Sci 8.53 (2016):2782-2787
Keywords : GARCH, Bandwidth, Cross validation and Kernel
Academic Editor : P. Sharmila
Conflict of Interest : None declared
Acknowledgements/Funding : The financial assistance received by the first author in the form of Senior Research Fellowship (PGS) from Indian Council of Agricultural Research (ICAR) is highly acknowledged.
Author Contribution : None declared
In terms of food and nutritional security, Potato is the crop, which is most important after Wheat, Maize and Rice in the World. The State West Bengal is the second leading producers of India. In this study -Area, Production and Yields of Potato in West Bengal for the period of 1963-2012 have been considered to apply different parametric models-linear, non-linear regression and time series models (Box Jenkins& GARCH); and nonparametric model. Suitable parametric model was selected on the basis of various goodness of fit criterion and assumptions of residuals. In case of nonparametric regression, optimum bandwidth was computed by method of cross validation. Here, epanechnikov-kernel was used as the weight function. Non parametric function was emerged as the one of the best fitted trend function among all selected models, wherein parametric models ARIMA (1,1,0) was identified for both area and production of Potato & ARIMA(1,1,1) was appropriate for Potato’s yield. Forecasting was made by selected parametric models up to 2020.