Models are generated in this research to describe and forecast the effect of global cocoa prices on burnished color production in Ghana utilizing a regression model accompanying time series mistakes. The problem of price airiness is particularly definite for Ghana due to her heavy reliance on cocoa exports for convertibility earnings. The experiment sought to decide whether the all-encompassing cocoa price can aid in expecting the future behavior of Ghana's cocoa result. Annual data from 1961 to 2010 were used to fit the model, accompanying out-of-sample data from 2011 and 2012. The reversion model with ARIMA(2,2,0) errors was considered the "best" model for the production changeable based on the results of abundant model adequacy techniques. The model was confirmed using the mean certain percentage mistake (MAPE) as forecast accuracy metric. Therefore, the 'best' reversion model's MAPE was 7.97%. The usual "best" ARIMA model, still, which was fitted to the result variable, anticipated a MAPE of 16%. This demonstrates that when the result variable was modeled in addition to the world price utilizing regression accompanying ARIMA errors, the MAPE was reduced. Therefore, a reversion model with ARIMA (2,2,0) mistakes is a more accurate mathematical method than the ARIMA method for envisioning Ghana's cocoa result.
Author(s) Details:
Sampson Ankrah,
Postgraduate Institute of Agriculture,
University of Peradeniya, Sri Lanka.
Kwadwo
A. Nyantakyi,
Postgraduate
Institute of Agriculture, University of Peradeniya, Sri Lanka.
E. Dadey,
SSNIT, Research Department, Accra, Ghana.
Please see the link here: https://stm.bookpi.org/EIAS-V4/article/view/10816
No comments:
Post a Comment