利用類神經網路建構物件導向量測雛型研究─以光電元件（TFT-LCD）電測製程為例

Abstract

目前高科技相關產業的電測製程中，電測軟體上所使用的測試程式多是有經驗工程師負責撰寫，不僅費時且常因重複試誤耗費過多的成本，在維護程式的程序上也常受工程師的更替所影響；本研究的目的即在利用類神經網路試圖找出歷史案例中符合目前製程所需的雛型測試程式，再藉物件導向技術系統化的處理，使得程式的維護及開發時間降低，更甚至於非程式人員也可撰寫程式。

The programs of testing software today are usually coded by experienced engineers. And it often cost a lot of time to build a testing software because of trying error and the relief of engineers. This case study are trying to use neural network to search the program prototype that fits the need of current process. And then use the concept of object-oriented programming to lower the cost of system development and maintanance.