Attribute Reduction Method Based on Generalized Grey Relational Analysis and Decision-Making Trial and Evaluation Laboratory

Abstract

Attribute reduction is a challenging issue in intelligent manufacturing. Existing methods are mainly based on rough set theory (RST) focusing on symbolic and discrete values. However, classical RST doesn't consider the complex interrelationship among conditional attributes and consecutive attribute values. Our article seeks to deal with this problem by proposing a hybrid framework based on generalized grey relational analysis (GGRA) and decision-making trial and evaluation laboratory (DEMATEL) method. GGRA is used to determine the initial importance of conditional attributes relative to the decision attribute and the direct relationship among conditional attributes. The causal relationships among conditional attributes are calculated by the DEMATEL method. Then we calculate the final importance of the conditional attribute relative to the decision attribute and apply a threshold to control the number of attributes entering the core set. Finally, an illustrative case and experiments are verified our method of attribute reduction with consecutive attribute values and complex interrelationship among attributes.