現今電子產業中，擁有自我品牌公司與代工組裝廠的合作模式為，代工組裝廠提供彈性靈活的供應鏈予自我品牌公司，供自我品牌公司在預測到市場變化時，可調整預測出貨量。而當調低預測出貨量時，即讓代工組裝廠面臨到呆滯材料的處理問題，而呆滯材料的處理對個案公司將會額外產生需要支付的HUB管理費用。 在景氣正常變動下，自由市場可發揮自行調節功能，然而當面臨到景氣循環波動大時，市場自我調節機制便難以發揮功效。本研究探討，在景氣循環波動大時，將嘗試以較保守的出貨量因子來預測未來出貨量。採用資料探勘(Data mining)的迴歸分析(Regression Analysis)模型，透過橫向與縱向角度來預測個案公司主力伺服器產品之出貨量。橫向面將透過其他產品出貨量來預測主力產品出貨量；縱向面將找出主力產品在新舊世代交替時，產品出貨量切換速率。 ;Current operation between ODM and OBM is, ODM will provide a flexible supply chain management operation to OBM. OBM can adjust their predicted product shipping volume once OBM foresees any change on product shipping volume. One the adjustment is to lower predicted product shipping volume, ODM will need to handle slow moving materials. It means, ODM will need to pay additional HUB storage fee/management fee. In a normal business cycle, market has its self-regulation system. When we face an abnormal business cycle, market self-regulation will be ineffective. This thesis intends to find a better solution to predict product shipping volumes under an unpredicted and abnormal business cycle. With data mining methodology, we will predict product shipping volumes with horizon view and vertical view by regression analysis model. In a horizon view, we will get a predicted product shipping volumes by other products shipping volume. In a vertical view, we will get an old-and-new product transition ratio when OBM needs to implement new generation product.
