This thesis deals with domestic flat-glass demand analysis and forecasting during the 4th Economic Development Plan period and present 3 sets of the flat-glass demand forecasting results obtained through various forecasting methods i.e. time series analysis, multiple regression analysis, and consumption sectors analysis. These results can be useful in establishing an economic planning for the flat-glass industry.
In time series analysis, two separate sets of forecasts under different assumptions on the seasonal demand are combined to yield forecasts with lower mean square error than either of the original forecasting technigues (WINTERS' MODEL and FOUR IER SERIES MODEL). Multiple regression analysis consits of two stages. In first stage, simple time series analysis is made to forecast independent variables selected by the UCLA-BMD Stepwise Regression Computer Program.
These forecasts are then used as inputs to a multiple regression model in order to forecast the flat-glass demand. In consumption sectors analysis, under the assumption that domestic demand for the flat-glass comes from 4 end-uses i.e residential new construction, nonresidential new construction, consumption repair and alteration, and nonconstruction, each sector is analized on the basis of the 1976 domestic glass sales and then demand required in each sector is forecasted.
The main conclusion of this study is that if the growth rate of the glass consuming industry increases significantly higher than the GNP growth rate(9.2%) projected in the 4th five-year Economic Development Plan, demand forecasts by the consumption sectors analysis are recommended. otherwise, forecasting results by the time series analysis and/or multiple regression analysis are considered to be preferable.