A Data Mining Model of Product Bug Reports and Its Application

• Main Authors: Chun-Chia Chang, 張純嘉
• Other Authors: Chein-Shung Hwang
• Language: zh-TW
• Published: 2010
• Online Access: http://ndltd.ncl.edu.tw/handle/83467944099101640289

碩士 === 中國文化大學 === 資訊管理研究所 === 98 === Influenced by globalization and the rapid development of technology, the contract manufacturing industry in Taiwan has gradually evolved from manufacture centric OEM to design centric ODM and is moving into OBM model. Because of this, the industry is aggressively developing its R&D skill. The trend of products nowadays is heading toward high precision, defect free, multifunction, and with multiple added value. In order to produce high quality products, high cost materials and accurate verification equipments are needed. High quality raw materials, high tech manufacturing equipments, sophisticated manufacturing process, and long manufacturing cycle are re-quired in the manufacturing process. If the manufacturing process encounters any failure, the irregularity of the product quality may incur huge losses. Therefore, con-trolling the product design and manufacturing quality has become a key to a successful product. In the lifecycle of a product, many documents and reports are generated from process of design to manufacturing to sales. Many knowledge and experience is hid-den inside this enormous database. Therefore, if we can efficiently utilize information and knowledge management, such as data mining, we can conclude many useful data to expedite R&D, customize production, and maintain high product quality. This research intends to connect the database among enterprise information systems. It summarizes bug reports into defect symptom, defect reason, and action using industrial database, decision trees, association rules, and neural networks. It analyzes failure reasons to prevent the occurrence of the same or similar issues to reduce manu-facturing time and the improve product quality.