Fuzzy modelling of patient state: distance-based versus rule-based approach

Two fuzzy approaches to medical decision making are discussed and applied. In the first method a fuzzy rule-based approach is used to define the state of a hypertensive patient. The second approach, based on the distances between fuzzy numbers representing norm, pathology and experimental observations, is applied to peripheral neuropathy electrodiagnosis in diabetic patients. The examples deal with conveying a precise description of patient state to the physician, as a decision aid. Both approaches define patient state as a continuous variable, are robust, appear not to depend on the shape of membership function, can be used with dependent parameters and may easily be extended to other cases of medical categorization