Study of TCM Diagnosis of Syndromes of Acute Exacerbation of Chronic Obstructive Pulmonary Disease Based on Dynamic Fuzzy Kohonen Network

Abstract

To explore common syndromes and their characteristics of acute exacerbation of the chronic obstructive pulmonary disease (AECOPD) based on dynamic fuzzy kohonen network.. By means of Fisher-iris data and Epidata software, information database, the study was established concerning patients with AECOPD , who were enrolled from four hospitals at 3 A -levels. The detailed procedures were as follows: (1) Selected kohonen net and fuzzy system, as the output of kohonen net can be reflect the graphical distribution characteristics of input samples; (2) On this base, a dynamic self-adaptive neural network was formed by increasing dynamic neurons; (3) guided by fuzzy theory, a dynamic neuro-fuzzy inference system was builded up with MATLAB6.5 software programming; (4) the model’s rationality was tested by Fisher-iris data and eventually obtaining characteristics of the common syndromes of AECOPD based on clinical data mining results and basic theories of traditional Chinese medicine. Through the rules conversion for the main and secondary symptom screening, nine syndromes and their corresponding main and secondary symptoms were determined. They included pattern of stagnated phlegm obstructing the lung, pattern of qi-yin deficiency of the lung and kidney, pattern of phlegm-dampness obstructing the lung, pattern of wind-cold attacking the lung, pattern of accumulated phlegm-heat in the lung, lung-kidney qi deficiency pattern, pattern of wind-heat invading the lung, lung-spleen qi deficiency pattern, pattern of exterior cold and interior heat. The pass rate reached 75.8.% approved by the data test on its rationality. The model can be used to study characteristics of TCM syndromes of AECOPD.