運用資料探勘技術於改善半導體測試業的計價系統效能

Abstract

在今日環境變化如此快速，製造的產品生命週期，相對跟著縮短，許多企業未因應此種局勢，都需要與客戶保持快速回應與緊密關係，且能快速提供附加價值產品與服務給客戶，相對才能比競爭者，更早一步提高公司內部快速反應機制，以因應客戶快速需求，提高競爭力。 在半導體測試業中，除了快速反應測試改變及測試結果外，測試計價的及時及整榷的提供也是很重要，也是提供企業決策的準則，並且滿足客戶的需求。 為達到此目的須改善目前計價系統方式，提供較彈性計價系統方式。此新的計價方式主要是應用資料倉儲整合觀念，彙整相關計價資料，另外再利用探勘資料規則方式找出計量行為，而算出計價量。再利用資料探勘的決策樹演算法，找出計價單價。新的計價方式即是運用這資料倉儲與資料探勘方式產生。. 本論文研究主要在改善目前計價系統運作方式，建立新的計價量及計價單價功能。這兩種新功能其效益不但能加速導入客戶的計價系統，並提高其快速建置的效能，特別是針對企業內部已存在而尚未導入計價系統客戶，另外亦能降低營收差異，縮短立帳作業時間，加速出口作業，提高出口效率，並能提供營收資料，以供整合至主管資訊系統(EIS)、決策支援系統(DSS)及商業智慧型系統(BI)，讓企業能即時做出整榷的決策。

In today's dynamic and changing environment, companies need to be responsive and closer to the customers, and deliver value-added products and services as quickly as possible. Companies also need to be able to support organizational information faster and better than their competitors. In semiconductor testing, beside emphasize fast reaction result in testing changed and result, billing’s real time and accurate provide data that is important also or too to provide the enterprise’s strategy guiding principal and satisfied customer requirement. To build the new intelligent billing system for this purpose, It major make use the Data Warehousing integration concept to consolidate these billing data and Decision Tree of Data Mining method to configured the billing rule. Both of data warehousing and data mining methods are to make up the intelligent billing system. The result of this paper is to design the new intelligent billing system. Speed up the implementation new billing system of customers, and improve the billing performance. Especially; the company have already existed many customers that didn’t stilled to implement the billing system. And finally this new intelligent billing system is to provide the accurate the almost real time revenue data that can view by customer, tester machine, and provide the real time revenue of enterprise to integrate the EIS , DSS or BI, let the enterprise to make a right strategic timely.