Multi-granularity dominance rough concept attribute reduction over hybrid information systems and its application in clinical decision-making

Abstract

Multi-granularity conceptual analysis has become an essential tool for data analysis and knowledge representation. This paper abstracts a class of uncertain decision-making problems, which integrate data and knowledge features from the clinical knowledge representation and decision-making process of integrated traditional Chinese medicine (TCM) and Western medicine. Under the framework of multi-granularity rough sets, we combine dominance rough sets with formal concept analysis methods to give a quantitative description of this uncertain decision-making problem. First, we define a multi-granularity formal context with attribute dominance relation. Then we give the concept lattice generation method under the formal context of multi-granularity dominance attributes and establish a multi-granularity formal concept analysis model that integrates data and knowledge representation. Meanwhile, the basic form of the multi-granularity dominance rough concept analysis model, the differences between existing rough sets and concept lattices, and their connections are discussed in details. Further, a clinical diagnosis decision-making model of integrated TCM and Western medicine, which fuses data driven Western medicine classification standard and knowledge-driven TCM diagnostic standard, is formed with the actual clinical diagnosis and treatment decision-making process of integrated TCM and Western medicine.