Table 6 clearly indicates the outstanding improvements using different measures such as MSAE (59% improvement), MAE (55% improvement), and MAPE (14% improvement)

Raw extraction · bibliography line

As illustrated in Figures 6-8 and the summary results reported in Table 6, the FSE model significantly improves the forecast accuracy when compared to Company B’s judgmentally adjusted forecasts. Table 6 clearly indicates the outstanding improvements using different measures such as MSAE (59% improvement), MAE (55% improvement), and MAPE (14% improvement). Figure 6: Absolute errors of forecasts compared to actual sales 0 50 100 150 200 105 110 115 120 Weeks Absolute Errors (Unit) Company B's Adjusted ForecastsFSE Model Forecasts 22 Figure 7: Relative errors of forecasts compared to actual sales Figure 8: Comparison of the two forecasts and actual sales in the test-set period 0 30 60 90 105 110 115 120 Weeks Relative Errors (%) Comapny B's Adjusted ForecastsFSE Model Forecasts 0 100 200 300 400 105 110 115 120 Weeks Demand (Units) Actual Sales Comapny B's Adjusted ForecastsFSE Model Forecasts 23 Table 6: Forecasting accuracy improvement in the test-set period Measure of Error Company B’s Adjusted Forecasts The FSE Model Forecasts Improvement in Accuracy MSAE 0.32 0.13 59% MAE 30.88 13.62 55% MAPE 48.60 41.65 14% 5 Conclusions In this paper, we propose a time series regression model