以貝氏信賴網路進行台灣電腦產業的策略分析及隨機模擬

Abstract

目前全球的電腦產業正面臨著許多機會與挑戰，像是全球PC電腦需求下降、各國紛紛提出政策試圖將製造業拉回國內生產或是客戶要求客製化等許多議題，而電腦產業中廠商發展的產品及服務不同，故本研究將電腦產業分為品牌電腦商、工業電腦商和代工電腦商三個子產業，又因台灣在三個子產業領域中都有發展良好且具有代表性的公司，故本研究以台灣三個子產業領域中的廠商為樣本進行分析。 本研究透過分類與迴歸樹(CART)找出三個子產業的重要變數，分析三個子產業價值鏈的不同之處，並利用貝氏信賴網路量化出變數之間互相影響的強度，最後利用模擬說明當自變數變動比自變數沒有變動時對依變數的機率影響程度，此舉可讓企業經營者明確的知道當他下決策時，對企業營利的確切影響為何。

Now, the global computer industry is facing many opportunities and challenges, such as the reduced demand of PC computers, country’s proposed policies, customized products and so on. Due to the computer industry manufacturers develop different way in their products and services, this study divided computer industry into computer brands, industrial computers and computer foundry of three sub-industries. Taiwan has well-developed and representative company, so we use Taiwan's manufacturers for the sample analysis. First, the study use CART (classification and regression tree) method to find out the important variables of the three sub-industries. Next, the study analyze the differences between the three sub-industry value chains. Third, the study use the bayesian belief network to quantify the strength between the variables. Finally, the study use the simulation to explain the independent variable impact on the dependent variable, so that business operators can clearly know what will happen on corporate profits when he made the decision.