Figure 1. Syntetos et al. Categorization Scheme. demand periods. The squared coefficient of variation represents variability of demand size. where N represents the number of periods having non-zero demand, ¢; represents the demand in period, ¢ represents the average demand considering only periods with non-zero demand (figure 1). Table 2. Descriptive statistics of the simulated data set. Table 1. Levels of the factors for each intermittent demand category. The steps included in the proposed method selection procedure are represented in figure 6. Table 3. Percentages for each intermittent demand category. Table 4. Descriptive statistics of the real data set. Step 2: In this step, data is categorized according to Syntetos ef al categorization scheme considering the Table 5. Performance results of forecasting methods for each type demand type. Table 6. Order of methods according to their performances for intermittent demand type. Table 7. Order of methods according to their performances for erratic demand type. Table 8. Order of methods according to their performances for lumpy demand type. Table 9. Order of methods according to their performances for smooth demand type. Table 10. Performance results of forecasting methods for each type demand type. Table 11. Order of methods according to their performances for intermittent demand type. Table 12. Order of methods according to their performances for erratic demand type.
