GWMA已被許多論文證實,GWMA與EWMA相較是相同的概念。但是,GWMA多增加一個參數,對較小波動偏移更加敏感,並且比EWMA獲得更加廣泛的應用。 然而,過去的研究往往被應用EWMA預測,但較少使用GWMA進行預測。這些論文評估預測性能使用指數加權移動平均法EWMA。然而,很少有應用GWMA進行預測。由於以往的研究亦證實,廣義加權移動平均法GWMA是一種簡單可靠的工具, Lin & Sheu 亦證實,GWMA比EWMA檢測製程中的變化更敏感,該GWMA方法可能是有用的預測。因此,使用GWMA預測是值得進一步探討。本研究試圖提供一種機器學習方法用以確定GWMA的雙參數。
Many papers on the GWMA confirmed that the GWMA concept is almost identical to the EWMA method. The GWMA method adds a parameter, is more sensitive to small fluctuation, and has a wider range of applications than EWMA. However, past research often found EWMA being applied to forecasts, but few used the GWMA method for forecasting. These papers used the EWMA method to assess forecast performance. However, few applications used the GWMA method for forecasting. Since prior studies confirmed that the EWMA forecast method was a simple and reliable tool, and Lin & Sheu also confirmed that the GWMA method is more sensitive and extensive than EWMA in detecting process shifts, the GWMA method could be useful for forecasting. Therefore, using the GWMA method in forecasting is worthy of further exploration. This study try to provide a machine learning method, and to determines the two parameters.