Performance evaluation of suppliers is increasingly recognized as a critical indicator in supply chain cooperation. Traditional performance evaluation methods have the problems of a simple buy/sell relation and in one’s subjective views between manufacturers and suppliers, and they lack objective automatic evaluation processes in the supply chain considered. Statistical techniques used for evaluation rely on the restrictive assumptions of linear separability, multivariate normality, and independence of the predictive variables. Unfortunately, many of the common models of performance evaluation of suppliers violate these assumptions. The study proposes an integrated model by combining K-means clustering, feature selection, and the decision tree method into a single evaluation model to assess the performance of suppliers and simultaneously tackles the above-mentioned shortcomings. The integrated model is illustrated with an empirical case study of a manufacturer for an original design manufacturer (ODM) to demonstrate the model performance. The experimental results indicate that the proposed method outperforms listed methods in terms of accuracy, and three redundant attributes can be eliminated from the empirical case. Furthermore, the extracted rules by the decision tree C4.5 algorithm form an automatic knowledge system for supplier performance evaluation.
