1Lecturer, Department of Agricultural Economics, Agribusiness and Extension, University of Energy and Natural Resources, Sunyani-Ghana
2Directorate of Prioritization, Monitoring and Evaluation, S.K. Rajasthan Agricultural University, Bikaner, 334006-Rajasthan, India
3Professor, Department of Agricultural Economics, S.K. Rajasthan Agricultural University, Bikaner-India
*E-mail: richard.bannor@uenr.edu.gh
Online published on 12 July, 2017.
This study modelled and forecasted wholesale monthly prices of coriander through the use of Autoregressive Fractionally Integrated Moving-Average Model (ARFIMA). Data on wholesale coriander prices from January 2003 to August 2015 in Kota market of Rajasthan were used. ARFIMA (1, 0.375, 2) was selected as the best fit model for forecasting of coriander prices based on minimum AIC and BIC values. However, the results showed that, the mean absolute percentage error (MAPE) of ARFIMA (1, 0.375, 2) on predictions from January 2005 to August 2015 was 7.21 whereas the mean absolute percentage error of ARFIMA (1, 0.375, 2) based on predictions from January 2003 to August was 7.08 percent. It is therefore recommended that even though ARFIMA model had a less MAPE, other forecasting models should be considered to forecast the prices of coriander in Kota and the best fit mode with least MAPE selected to promote marketing and price discovery and also improve the bargaining power of coriander producers and other value chain actors in Rajasthan.
ARFIMA, coriander, long memory, marketing intelligence, Rajasthan, wholesale price
