摘要 近幾年來,薄膜電晶體液晶顯示器(TFT-LCD)技術逐漸取代傳統冷陰極射線管(CRT),成為新的顯示技術,而此新的顯示技術被廣泛運用於手機、數位相機、個人數位助理(PDA)、桌上型顯示器、筆記型電腦及液晶電視,成為3C市場上最閃耀的產品。為了能對液晶顯示器的產品生命週期進行有效地預測,我們使用一般化Bass擴散模型 (Generalized Bass Model) (Bass; Kriahnan and Jain,1994)來估計Bass Model 的創新係數與模仿係數,其估計結果可以幫助我們整理液晶顯示器在:(1)未來每個期間的銷售狀況、 (2) 在尖峰期間的銷售量以及 、(3) 該產業何時會到達成熟期。本研究採用的資料來自Display Search、MIC、拓墣產業研究所等,這些資料在本研究中都被用來描述使用一般化Bass擴散模型的好處。當整個銷售歷史被納入考量時,由一般化Bass擴散模型獲得的參數估計與最小平方法 (Ordinary Least Squares , OLS) (Bass,1969) 、最大概似估計法 (Maximum Likelihood Estimation, MLE) (Schmittlein & Mahajan, 1982) 非線性最小平方法 (Non-linear Least Squares , NLS) (Srinivasan & Mason, 1986) 相較而言,顯示出好的一致性,這個結果與 Venkatesan; Krishnan and Kumar (2004) 所做的實証結果一致。當可獲得的資料點一直持續到反曲點之前能被使用時,一般化Bass擴散模型對未來銷售、尖峰銷售時期與尖峰時期銷售量估計與現今可得的測量技術所估計的數值相比,提供了較佳的預測。
ABSTRACT In the last few years,The thin film transistor liquid-crystal display (TFT-LCD) technology substitutes for traditional cold cathode ray tube (CRT) gradually and becomes the new demonstration technology. This new technology are application of Mobile Phone, Digital Camera, PDA, Desktop Display, Notebook as the popular product in the 3C market. In order to carry on to the liquid-crystal display product life cycle forecast effectively, we use the Generalized Bass Model (Bass; Kriahnan and Jain,1994) to estimate the Innovation coefficient and imitation coefficient of Bass Model. This estimation result can help us to figure out the LCD monitor (1) The sales status of each period in the future.(2) The sales quantity in the peak period.(3) When will this industry reach the maturity? When we consider the sales history of LCD monitor, the coefficient estimation from Bass Diffusion Model is better than Ordinary Least Squares , OLS (Bass,1969), Maximum Likelihood Estimation, MLE (Schmittlein & Mahajan, 1982) (Non-linear Least Squares , NLS) (Srinivasan & Mason, 1986) and same as the result from Venkatesan; Krishnan and Kumar (2004) When the data we obtain can be adapted till the reverse point, we found the Bass Model provide the best forecasting on the sales and the quantity on peak period, compare with the data from Generalized Bass Model and the measure technology,