Research on diagnostic models of pneumonia syndromes in TCM based on fuzzy-neural net

To explore methods of establishing standard models of pneumonia syndromes in TCM (traditional Chinese medicine) by studying the results of data mining of pneumonia. First, in accordance with the selection criterion, 1058 pieces of clinical data from the patients with pneumonia were collected by clinical epidemiological methods. Secondly, fuzzy neural net models were built up on the basis of dynamic kohonen network and their reliability was tested with the Fisher-iris data. Then, with the help of the models the clinical data was studied and the diagnostic criterion for commonly-seen syndromes of pneumonia was obtained according to TCM basic theories. Simultaneously, the reliability was tested by data check-up. The coincident diagnostic rate reached 86% in comparison of the diagnostic criterion and original diagnostic data. The model, for its rational characteristics, can be applied to the study of diagnostic criterion for pneumonia syndromes.