以資料探勘技術提升覆晶構裝技術之產品開發能力 ; The Analysis of Stock Factor For automotive manufacturing by using Data mining Technique

NCUIR > School of Management at National Central University > Executive Master of Industrial Management > Electronic Thesis & Dissertation > Item 987654321/44023

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/44023

Title:	以資料探勘技術提升覆晶構裝技術之產品開發能力;The Analysis of Stock Factor For automotive manufacturing by using Data mining Technique
Authors:	鄧仁棋;JEN-CHI TENG
Contributors:	工業管理研究所碩士在職專班
Keywords:	可靠度試驗;覆晶構裝技術;決策樹;資料探勘;decision tree;data mining;flip chip packaging technology;reliability test
Date:	2010-07-07
Issue Date:	2010-12-08 14:46:09 (UTC+8)
Publisher:	國立中央大學
Abstract:	隨著科技發展、生活水準提高，兼具高功能、輕、薄、短、小的消費性產品的需求也逐日增加，驅使電子製造技術不斷發展演進，而構裝技術也面臨更大的挑戰，電子元件構裝公司需要投入更多的研發成本、與時間，去克服先進技術之電子元件所帶來的問題，以符合電子產品之需求與功能。為了降低研發成本、與減少產品開發的時間，需要有效的方法提升研發人員產品開發的能力，進而達成此目標。本研究即是利用資料探勘技術，以決策樹、類神經網路、以及邏輯斯迴歸分析方式，藉由電子構裝產品過去研發的經驗，以構裝產品的結構、製程條件、與使用的材料做為預測變數，建構出構裝產品可靠度試驗結果的預測模型，並找出較準確的預測模型，再以模型之預測能力做為研發人員評估新條件的參考依據，並利用模型找出影響試驗結果的重要變數，進而找到變數的最佳水準。以期透過模型的幫助，增加人員效率與降低人力、物料成本，以達到降低研發成本與減少產品開發時間的成效，提升產品開發能力與企業競爭力。With advances in technology, and enhancement of the living level, the demand for multi-functional, thin, and light-weighted products increases day by day. It drives the electronic manufacturing technology to improve continuously, and also makes packaging technology face great challenges. Packaging companies of the electronic components require lot of money to invest in R & D, and time to resolve quality issues arising from the new electronic components with advanced technology. In order to reduce the development cost and the time of the product development, effective methodology is needed to enhance the ability of product development of R & D. Based on our experiences on the electronic packaging products, this study uses the techniques of data mining with consideration of product structures, process conditions, and materials of the packaging products to build prediction models, decision tree, neural network, and logistic regression models, of reliability tests of packaging products. Through this study, we will find more accurate prediction models, uses these models to find the important variables that influence the test result most, and then find the optimal level of these variables. This study will be able to help to increase staff efficiency and reduce the cost of human and material to meet the effect of reducing the cost in R & D and the time of product development, and enhance the ability of the product development and the competitiveness of the enterprise.
Appears in Collections:	[Executive Master of Industrial Management] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	916	View/Open

社群 sharing

Loading...