Applying Data Mining Techniques to Product Abnormal Detection for Semiconductor Package Manufacture

• Main Authors: Che-kun Shih, 史哲坤
• Other Authors: leorean
• Format: Others
• Language: zh-TW
• Published: 2005
• Online Access: http://ndltd.ncl.edu.tw/handle/65692185167163250676

碩士 === 義守大學 === 資訊管理學系碩士班 === 93 === The semiconductor package manufacture enterpriser is always caring about the questions of new process yielding good result while in the course of production. Because the yield will be concerned the quality of products and the profitability. It must invest huge fund and technology, so that the unit cost of products became very expensive. For this reason, we must pay attention to the quality management of process. But all products abnormal questions are the unchangeable fact; Afterward the holding lot measures or root cause analysis and improvement will be only remedial measures. If we can use large amount of materials storage which noting down in the data warehouse and use the materials in the data mining to analyze by technological method, then finding out the products abnormal in advance will help to improve yield and reduce production cost. According to our knowledge analysis can sum up the reasons that cause of product abnormal. For example, the associations of some materials are very apt to cause the breakage. So it is quite important to select the proper materials to match; besides in a specific period of some production units appear the high rate of the artificial mistakes. Then the company may consider about the system of education and training of improving the staff.