本研究利用灰色預測模型對個案公司2016年1月至2017年3月槍嘴產品銷售資料進行建模與模擬預測，並將預測結果進行殘差的評析，以驗證模型之預測效能。研究結果顯示，GM(1,1)滾動模型對產品需求預測的效果較佳，尤以資料數目4筆的GM(1,1)滾動模型預測與驗證效果最佳，預測平均準確度92.83%，驗證準確度也都高於90.0%，並由表4-4的分析與比較結果得知，GM(1,1)滾動模型預測2016年1月至2017年6月之銷售資料，其平均預測精度為93.10%優於個案公司以經驗值預估的精確度77.82%。因此，GM(1,1)滾動模型可作為本研究個案公司產品需求預測之參考。;This study uses the grey prediction model to model and simulate the sales data of the products from January 2016 to March 2017. The prediction results are evaluated by residuals to verify the prediction performance of the model. The research results show that the GM(1,1) rolling model has a better effect on product demand forecasting, especially the GM(1,1) rolling model with 4 data points has the best prediction and verification effectiveness, and the average prediction accuracy is 92.83%. The verification accuracy is also higher than 90.0%, and from the analysis and comparison results in Table 4-4, the GM(1,1) rolling model predicts sales data from January 2016 to June 2017, and its average forecast accuracy, 93.10% ,was better than that of the case company with an accuracy of 77.82% by using the experience method. Therefore, the GM(1,1) rolling model can be used as a reference for the product demand forecast of the P company in this study.
