Abstract:
The purpose of this research was to compare the errors between combined and individual forecasts by using weighted average of Newbold and Granger and 8 individual methods. These 8 methods were SMA, SES, DES, LES, TES, HWS, REG and B-J, using 2 error measures MAPE and MSE as criteria. เท addition, the researcher forecasted 10 lead times. Two data bases were used in this study. The first were three sets of 65 monthly time series data with seasonal variation, namely Thai books, English books and theses เก circulation, obtaining from the Educatonal Information Center, Faculty of Education, Chulalongkorn University. The second were two sets of 14 yearly time series data with nonseasonal variation of Science and Mathematics teachers collecting from Planing Division, Department of General Education. The research instruments were data recording forms. The data were analyzed using graph and regression analysis to check for trends and seasonal variation, using individual and combined forecasts to check the forecasting results and lead time forecasting. The research findings were as follows : 1. The numbers of Thai books, English books and theses checking out from the library showed quadratic and seasonal variation, linear and seasonal variation, and linear and seasonal variation with the additive integrated models at .01 significance level. Furthermore, the numbers of Science and Mathematics teachers showed quadratic trends at .01 significance level. 2. The best methods for time series data with seasonal variation, Thai books and theses, were the combined forecast that came from HWS and REG using the second and first weighted average. The best method for English books was HWS. The best methods for time series data with nonseasonal variation, Science and Mathematics teachers, were the combined forecast that came from combining every methods together (SMA, SES, DES, LES, TES and REG) using the first and second weighted average. 3. The forecasts from November 1999 to August 2000 for Thai books would be 4,786, 4,554, 4,868, 5,091, 2,869, 2,811, 2,290, 4,695, 5,321 and 5,751 1 for English books would be 789, 718, 723, 710, 448, 433, 379, 750, 771 and 821, and for Theses would be 9,740, 8,057, 8,754, 8,833, 8,078, 6,605, 4,254, 7,372, 8,155 and 9,091, respectively. The forecasts of the number of teachers in the academic year of 1999 to 2008 for Science teachers would be 18,189, 18,714, 19,251, 19,801, 20,362, 20,937, 21,523, 22,122, 22,733 and 23,356 and for Mathematics teachers would be 12,637, 13,239, 13,887, 14,581, 15,320, 16,104, 16,934, 17,809, 18,730 and 19,696, respectively.