Prospects of Purchasing—An Evaluation Model for Data Mining Approaches for Preventive Quality Assurance

This contribution conceptualizes an evaluation model for data mining approaches for preventive quality assurance in purchasing. Future purchasing heavily relies on data analytics and purchasers need to be equipped with suitable tools and skills. A major prerequisite is to collect the respective data and to apply suitable algorithms to generate added value. This will free human capacity for more strategic initiatives and will provide an increase in flexibility and productivity within the company. To derive a valid evaluation model, data mining methods apt for preventive quality assurance being a binary classification problem are presented. Based on a seven steps data mining approach and a literature analysis, requirements for the data mining methods are derived. Subsequently, the criteria are aligned with exigencies in purchasing. This leads to a weighting of the respective criteria. The application of the evaluation model shows that support vector machines and k-nearest neighbors seem to be the best suitable data mining methods for preventive quality assurance in purchasing.