Applying The Neural Fuzzy System To Predict Sugar Consumption And Production: Evidence From The Egyptian Sugar Industry

摘要

The current study addresses the prediction of both consumption and production of sugar. The former function's dependent variable is sugar consumption quantity (in kiloton KT) and the independent variables are the price index number, production quantity (in KT), per capita income, number of population and import quantities (in KT). The latter function's dependent variable is sugar production (in KT); and the independent variables are raw sugar cane, sucrose ratio, agriculture area (in thousand acres), average of productivity per acres, working days in the manufactured and impurities ratio. A secondary data sources (i.e. a 23-years time series) were utilized for this purpose. Evidence from the Egyptian Sugar industry revealed that, Neuro-Fuzzy technique gives the minimum value of MSE. The attribute of this paper is the ANN can produce accurate predictions for the low noise data and the neuro-fuzzy model can produce more accurate predictions for the highly noise data.