|
Objective: To study the establishment and application of the standard model of type 2 diabetes syndrome diagnosis as a starting point to explore the syndrome diagnostic criteria established. Methods: First, from the Internet to collect the Fisher-iris data; Secondly, computer literature search to retrieve, collect the from 1984 to 2005 Nian between the Chinese bio-medical literature CD-ROM databases and Chinese journals full-text databases such as literature data; through the issuance of clinical investigation table, the selection criteria of the First Affiliated Hospital of Henan College of three hospitals from 2003 to 2006 in line with the cases of outpatient and inpatient clinical investigation in patients with type 2 diabetes. Data preprocessing, and the establishment of a literature database and clinical databases. Select the artificial neural network (Artificial Neural Network, ANN), fuzzy system (fuzzy system, FS), to carry out the study of the establishment and application of the standard model of type 2 diabetes syndrome diagnosis. The programming with MATLAB6.5. This thesis uses the dynamic Ke Helun network, its output results can reflect the the graphical distribution characteristics of the input sample, the right adjustment of the links, so that the probability density distribution of the right of the input sample distribution is similar, and on this basis, by increasing dynamic neurons form a dynamic adaptive neural network. Be dynamic layer neurons stable weights between the input layer and dynamic layer attributes of the corresponding rules of the fuzzy inference system membership function center, compared with a neural network recognition rate, and constantly adjust the fuzzy rules and the corresponding function parameters, the final optimal fuzzy rules. Model to test its reliability by Fisher-iris data, based on the results of clinical data mining, Reference data mining results, according to the basic theory of TCM, type 2 diabetes common syndrome diagnostic criteria, and to examine the suitability of the test data. Results: Based on the Fisher-iris data stable dynamic layer neurons increased to nine, get three fuzzy rules, the test sample test result identification rate of 94%. The model used in the literature data, the stability of the dynamic layer neurons increased to 22, get nine fuzzy rules, the use of the test sample test result identification rate of 86%. Based on the results of the clinical data is stable dynamic layer neurons increased to 118, the number of fuzzy rules to obtain 24, 74% of the test samples the recognition rate of clinical data. Converted by the rules, primary and secondary disease screening, according to the the common syndrome discriminant set and TCM syndrome dialectical standards, clear six witnesses and their corresponding primary and secondary symptoms 6 syndrome type were Qiyinliangxu the certificate the stasis syndrome Dryness Tianjin injured permits, stomach heat flaming certificate, hot and humid, retention syndrome, kidney yin deficiency syndrome. Common syndrome in type 2 diabetes diagnostic criteria: 1) Qi Deficiency: Main symptoms: malaise, fatigue, red tongue, palpitations, thin moss. Secondary symptoms: thirst, polydipsia, rapid pulse, spontaneous, less moss, insomnia, five upset hot, dry mouth, and they dry, shortness of breath, and sweating. 2) kidney yin deficiency syndrome: the main symptoms: frequent urination urine, red tongue, weak waist, Mi Gan grease. Secondary symptoms: little coating, rapid pulse, dysphoria heat, thirst, tinnitus and deafness, dry mouth, insomnia and more dreams. 3) stasis syndrome: the main symptoms: petechiae tongue, dark complexion, pulse string, Di dark tongue. Secondary symptoms: dark red tongue, numbness, limb pain. 4) swollen dry Jin injury Card: Main symptoms: thirst, polydipsia, red tongue, throat, dry mouth, tongue yellow, urine and more frequent urination. Secondary symptoms: thin fur, eat easy to hunger. 4) stomach heat flaming certificate: Main symptoms: easy to eat more hunger, yellow tongue coating, easy thirsty to drink, red tongue. Secondary symptoms: fever, irritability, palpitations, urine volume. 5) damp heat retention syndrome: the main symptoms: epigastric fullness, greasy moss, yellow tongue coating, abdominal fullness, body weight difficulties. Secondary symptoms: red tongue, tinnitus and deafness, bitter mouth, thirst, polydipsia, sticky mouth, dry mouth. Conclusion: Fisher-iris data to test the model shows that the method for fuzzy classification rules reflect higher precision the learning sample concentrate regularity, indicating the reliability of the model. By comparative literature data model based mining results, with the four card type based on the results of the clinical data obtained in Qi Deficiency, blood stasis, Dryness Jin injury certificate, Kidney Yin Deficiency corresponding primary and secondary symptoms basic The mining results indicate that the model can be used for the study of type 2 diabetes syndrome diagnostic criteria.
|