Research On The Technology Of Decision Support Of Tibetan Medicine Treatment Of The Common Diseases Of The Plateau Based On Medical Data Mining

Tibetan medicine is an important part in our traditional national medicine and pharmacy,and has a unique theoretical system and clinical curative effect.Because of the weak base of clinical research,the subjective factors of doctors are strong in the classification and diagnosis of disease syndrome,with lack of standardized and unified scientific diagnosis and treatment system.Using data mining technology to study Tibetan medical decision-making is the main contents of methods and means that inheritance of Tibetan medicine and prevention and treatment of disease.In this paper,firstly,the clustering algorithm was used to classify the common diseases of the plateau(chronic atrophic gastritis).The train of thought of Tibetan medicine syndrome type classification research was proposed which combined with clinical diagnosis and treatment data.The two step clustering and K-means algorithm were used to classify the common diseases of the plateau(chronic atrophic gastritis)as a preliminary classification,and then the Gower distance +improved k-modes method was proposed,which was verified.The final determination of the common diseases of the plateau(chronic atrophic gastritis))should be divided into four syndrome types.The objectivity and standardization of the Tibetan medicine syndrome type was achieved.Correlation mining of symptoms of disease were conducted through association rules.In view of the research content and the shortcomings of Apriori algorithm,improved Apriori algorithm was proposed based on constraint thought.The core is to constrain the frequent itemsets,only keep the items that conform to the constraints,and carry out subsequent operations.It not only shortens the running time,but also makes the mining rule set more practical,which reflects the efficiency and feasibility of mining.Finally,according to the clustering results of syndromes,Tibetan medicine syndrome classification model was founded by using classification algorithm.Based on the grey box method,the prediction model was established by using classical classification algorithms,such as Naive Bias,Logistic regression,neural network and random forest.Multiple Evaluation Criteria,such as prediction accuracy,Kappa Statistics,TP Rate,ROC Area were evaluated in order to find an ideal modeling method.The last,based on Naive Bayesian algorithm of grey box method,the common diseases of the plateau(chronic atrophic gastritis)of prediction model of Tibetan medicine treatment was realized,the prediction accuracy rate reached 87.2%.