Double-quantitative multi-granularity kernel fuzzy rough sets model and its application in rheumatoid arthritis risk assessment

The medical big data of combined Chinese and western medicine diagnosis and treatment (CCWMDT) is difficult to be used effectively in clinical medical decision-making because of its complex structure and multi-source storage. In this paper, we discuss the problem of multi-attribute group decision making (MAGDM) in complex information systems with multi-source, diversified and hybrid attribute information. With the help of traditional information systems, a hybrid diversified attribute information system (HDAIS) is first defined. Further, with the involvement of kernel functions, a binary relation over HDAISs is then presented. At the same time, in response to the problem of inaccurate information, a double-quantitative model is introduced to extract effective information from the perspective of absolute and relative metrics, and the rough approximations of a decision target concept under the multi-granularity framework are given, namely, a double-quantitative multi-granularity kernel fuzzy rough set (DQ-MGKFRS). Based on the above, a group decision method is established on DQ-MGKFRS to realize the information fusion of HDAIS. Finally, the proposed method used in the medical decision problem of rheumatoid arthritis (RA) diagnosis. The results of comparative analysis and sensitivity analysis illustrate the feasibility, effectiveness and robustness of our method.