貨損類型與嚴重程度分模式建構-以電子產品為例

Abstract

供應鏈風險管理一直是近年來非常熱門的議題，而在供應鏈物流活動當中經常發生的貨損事件，是造成延遲交貨甚至是供應鏈中斷的原因之一，但在實務或學術文獻卻鮮少有針對全球運籌企業各種運輸方式造成的貨損類型與其損防建議進行深入研究。因此，本論文將以從事全球運籌之電子產業為例，蒐集並分析電子產品在頻繁的全球物流活動當中發生的所有貨損案件，再運用資料探勘技術中常見的決策樹分析方法，發展出電子產品之貨損類型與嚴重程度分類模式，並驗證貨損分類模式之有效性與實務管理之參考價值。 本論文之研究架構是參考知識發掘流程，並於資料探勘步驟運用決策樹分析方法產生貨損分類模式。決策樹分析方法產生的貨損分類模式，能提供實務管理者容易解讀之分類規則與物流條件組合，做為預測貨損結果之參考工具，因此能協助管理者建立出有效的貨損防阻計畫。此外，本研究使用的知識發掘與資料探勘技術，也期望能夠推廣至不同的產業與產品，用於企業物流活動之貨損分析與貨損防阻之管理。

Supply chain risk management has been a popular topic in recent years. Cargo loss in supply chain and logistics activities has been the major cause of delays and supply chain disruption; however, rarely do academic papers or examples in practice provide comprehensive studies focusing on types of cargo loss and loss prevention in various modes of transportation used by global companies. Therefore, this paper will mainly emphasize on the global electronics industry, gathering and analyzing all causes of its cargo losses in transit. Decision tree analysis, as generally used as one of the data mining technics, will be adapted to develop classification models for cargo loss type and severity. This paper also examines the effectiveness and value of these models in practice. The frame work of the study is based on the process of KDD (Knowledge Discovery from Database) and the classification models are produced by using Decision Tree Analysis in the step of data mining. The cargo loss classification model produced by the decision tree analysis helps the manager understand the classification rule and the combination of logistics conditions easily so as to become a reference tool to the result of cargo loss prediction and helps the manager to make a more effective plan on cargo loss prevention. Furthermore, the KDD and Decision Tree Analysis used in this study are expected to be spread out in different industries and products, to manage their cargo loss analysis and cargo loss prevention in the logistics activities.