The Application Of Data Mining In The Supplier Evaluation

Supply Chain Management is a kind of new management concept, and is given more and more attention in domestic and abroad. It is regarded as one of the foundation of improving enterprise competitive power and establishing trustful relationship with up-chain and down-chain enterprises. Supplier evaluation evaluates up-chain enterprise performance in the supply chain. It plays a very important role in selecting suppliers, inspiriting supply chain partners and improving supply chain performance. At present, the fact is that there is not a precise and objective supplier evaluation system in domestic enterprises, which has greatly weakened the performance of the supply chain.This thesis relies on the database of enterprise management information system. By analyzing enterprise supply data using data mining methods, it classifies suppliers using neural network, in order to evaluate suppliers. The main research is included as following:(1) The difficulties of supplier evaluation and disadvantages of existing supplier evaluation methods have been stated. A data mining based supplier evaluation model has been proposed. The model evaluates suppliers by analyzing data in the database of the enterprise information system, so that man-made disturbances are greatly eliminated. The model has been applied to fuel supplier evaluation, and the experiment showed its high accurate rate.(2) Through the experiment, the superiority of kernel covering algorithm comparing with other algorithms has been analyzed. The experiment showed that the accurate rate of kernel covering algorithm is higher than that of alternative covering algorithm and SVM (Support Vector Machine). Also, the network training of kernel covering algorithm is faster than SVM, and the classifier produced by kernel covering algorithm is simpler than that produced by alternative covering algorithm. So kernel covering algorithm is more applicable for supplier evaluation.(3) Fuel Supplier Management and Evaluation System has been designed and developed using rapid prototyping method. Rapid prototyping method greatly improved the communication quality between developers and users during the development process, shortened the development cycle and enhanced the satisfaction of users.The research in this thesis is a new application of data mining technology. It provides a new thought of supplier evaluation. The supplier management system enables enterprises to evaluate suppliers scientifically and rationally. It provides credible decision-making evidence for enterprise managers.