An improved fuzzy connectivity diagnosis method

The connectivity diagnosis method proposed by Norris et al (1987) produced disappointing results in our experiments. In exploring the reasons for this, we designed a new algorithm in the same spirit as the original, and which is both more intuitive and more effective. By making use of the frequency of occurrence of symptoms in the training set, the new algorithm is better at selecting representative syndromes, and can recognize more than one syndrome for each disease.<>