透過您的圖書館登入
IP:18.218.184.214
  • 學位論文

應用灰色支援向量自我迴歸於新產品之需求銷售預測-以iPhone為例

Demand Forecasting Analysis for iPhone Using Grey Support Vector Auto Regression Model

指導教授 : 曹銳勤

摘要


因iPhone手機熱賣,打開了一股智慧型手機的風潮,智慧型手機市場的無限商機且市場的發展快速,使得iPhone手機的銷售需求預測,對於相關產業及蘋果公司的市場評估、投資經營和經營管理都具有相當的必要性。 由於iPhone新上市產品的銷售量數據資料少且不確定,因此本研究採用灰色理論分析適用於少數不確定問題的優點及支援向量迴歸中支援向量及容忍誤差的分析概念,提出利用結合灰色預測和支援向量迴歸兩者優點的模型“灰色支援向量迴歸模式”並結合自我向量迴歸的概念,預測新產品的銷售需求,希望可以藉此提高預測的精準度;並以MAPE及本研究所提出方法的作為預測績效衡量指標。實證結果顯示AutoGSVR模式(MAPE=11.69%)表現優於GSVR模式(MAPE=12.99%)、簡單迴歸模式(MAPE=18.06%)、指數平滑模式(MAPE=27.38%)及GM(1,1)模式(MAPE=13.06%),並可成功外插預測2011年第二季的iPhone手機銷售數字,MAPE值可達0.01%。

並列摘要


The surprising boom of iPhone series has led the trend of smartphones. There are infinite business opportunities in the smartphone market and the market grows so fast, so it is significant for us to keep an eye on the development of smartphone market. The forecast of iPhone demand is necessary in the market assessment, investment and management of relevant industries and Apple Computer, Inc. Because the iPhone new going on the market product's sales volume data information are few, and is indefinite, therefore this research uses the grey theory analysis to be suitable in the minority indefinite question merit and in the support vector return supports the vector and the tolerance error's analysis concept, proposed that forecasts and the support vector using the union gray returns both merit the model “the gray support vector return pattern” and unifies the self-vector return the concept, forecast that the new product the sales demand, hoped may take advantage of the enhancement forecast accurate; And proposes the method achievement forecast achievements weight target by MAPE and this research institute. The real diagnosis result showed that the AutoGSVR model (MAPE=11.69%) showed significantly better than GSVR model (MAPE=12.99%), simple regression model (MAPE=18.06%), single exponential smoothing model (MAPE=27.38%) and GM(1,1) model (MAPE=13.06%), and can successfully extrapolation forecast that in 2011 the second season the iPhone handset sale digit, the MAPE values up to 0.01%.

參考文獻


技術研究所,1997。
15. 張逸凡,支撐向量機在即時河川水位預報之應用,國立成功大學水利及海洋
24. 歐宗殷,資料探勘為基礎之零售業銷售預測模式-以連鎖超商鮮食商品為例,
18. 陳寬裕,演化式支援向量迴歸於旅遊需求之預測,私立長榮大學經營管理研
22. 詹淑鳳,應用灰色支援向量迴歸預測國際旅遊需求,碩士論文,私立淡江大學管理科學研究所,2010。

延伸閱讀