Application of Neural Network to Supplier Selection

• Main Authors: Bo-Ying Tseng, 曾柏穎
• Other Authors: 林金賢
• Format: Others
• Language: zh-TW
• Online Access: http://ndltd.ncl.edu.tw/handle/79413249242791606080

碩士 === 國立中興大學 === 企業管理學系所 === 96 === In the trend of globalization for businesses, past production methods are not suitable for the environment today. In recent years, purchasing for businesses has become a huge component of operating cost. For this particular reason, the purchasing activities of businesses have evolved from day to day routine procedures to strategic purchasing procedures. Issues regarding purchasing and the selection of suppliers are related to one another, therefore as the importance of purchasing rises, the selection of suppliers become an important issue as well. Purchasing satisfaction is an important performance evaluation index for all businesses. In the past, the evaluation of suppliers is based on performance scores, but not purchasing satisfaction after the purchasing procedure.Hence, after reviewing several related literatures, this research discovered that neural network is the best among the selection methods. Neural network can solve complicated non-linear problems in terms of supplier selection. This research uses neural network as a method in assessing suppliers. Furthermore, this research uses company X as the subject of this case study, using the company’s order information as sample to construct a prediction model of supplier selection. This model uses criteria of supplier selection as independent variables, and different from past approaches, uses purchasing satisfaction as the dependent variable. This research also uses a regression model to compare the accurateness of the predictability of models. The result of this research had indicated that the neural network model can indeed provide a better predictability and can also bring forward a better model of making supplier selection for businesses.