Using classification techniques detect the new apparel items of sales pattern -Taking E company as an Example.

• Main Authors: Li, Che-Hao, 李哲豪
• Other Authors: Lee, Rong-Gui
• Format: Others
• Language: zh-TW
• Published: 2018
• Online Access: http://ndltd.ncl.edu.tw/handle/535rz7

碩士 === 國立交通大學 === 工業工程與管理系所 === 106 === The apparel retail industry has highly uncertain, first of the uncertain of fashion trend and the other is uncertain of demand. Especially the apparel life cycle relatively shorter than the other consumer goods. In the modern society of professional job assignment, the apparel supply chain can be divided into fiber, weaving, dyeing and finishing manufacturing. Under the influence of procurement, production, and retail globalization, When the decision maker of apparel retailer making a wrong decision, that will be prone to out of goods or increase the slow-moving goods. Therefore, this study will analysis on the sales data of new and old apparel goods, It will make to increase, reduce or stop the new goods by sales performance of new goods, even recycle and redesign the new. Enterprises resources will put in the goods with strong sales performance, and it will be possible to achieve the goal of apparel retailer good selling goods are not out of goods and decrease the slow-moving goods of weak sales performance. The clustering results show that the group one, two, and three represent the products' sales performance is excellent, generally and weak. The conclusions show that the accuracy of detecting the sales ability of new products is about 28 of 38 new products are as expected. Accurate rate about 77.78%. However, most of the goods are concentrated in the group with the general sales performance. The follow-up study should upgrade the group to 4 groups. It may be easier to see the changes in the sales ability of new products, which should help improve the accuracy of detection.Keyword：Apparel Industry, Clustering Analysis, Distance Matrix.