Abstract:
The present study was carried out at statistical Section, Ayub Agricultural Research Institute, Faisalabad, Pakistan during the year 2013 to examine the production and productivity trends of wheat in Pakistan for the period 1947-48 to 2012-13. Several nonlinear models are applied to fit the trends. Most suitable model selected on the basis of maximum Adjusted R square, relatively minimum values of mean absolute error, root mean squared error and normally, independently, identically distributed error term. None of the parametric models is selected to fit trend of area sown under wheat in Pakistan, nonparametric regression model with kernel smoothing was employed. Sinusoidal model was selected as a suitable model for the both production and productivity of wheat in Pakistan as it fits the trends with 99 percent predictability in case of production and 98 percent for productivity alongwith all the satisfactory assumptions regarding error term. The root mean squared error of Sinusoidal model for production (0.779) and productivity (99.58) are relatively lower than all other models compared in the study to fit the trends. Mean absolute error also observed minimum in case of Sinusoidal model for both the production and productivity with values 0.623 and 84.778 respectively. The rate of yearly change due to suggested model is 2.95 percent in case of production and 1.89 percent for productivity.
Page(s):
115-124
DOI:
DOI not available
Published:
Journal: Journal of Agricultural Research, Volume: 55, Issue: 1, Year: 2017
Keywords:
Root Mean Squared Error
,
Mean Absolute Error
,
Adjusted R square
,
Sinusoidal model
,
Nonlinear models