Research On The Method Of TCM Diagnosis And Treatment Based On Supervised Learning Classification Algorithm

Multi label classification is an important branch of data mining.Because of its wide application field,it is also a research hotspot in the era of big data.There is no consensus on syndrome differentiation and treatment of traditional Chinese medicine.Each family is still based on experience and has strong subjectivity.In order to deal with the doctor's diagnosis experience scientifically,the diagnosis and treatment data are trained into an objective diagnosis model,and the multi label classification technology is applied to the research of TCM insomnia syndrome differentiation.It is hoped that this method can help clinical doctors and highlight its value.Due to the heterogeneity,diversity and redundancy of the medical records of traditional Chinese medicine,in order to describe patients' symptoms objectively and fairly,it is necessary to classify and quantify the information of the medical records and record it according to the multi label data specifications for training the algorithm model.The algorithm model adopts the multi label classification model of the improved KNN algorithm,and discusses the multi label algorithm,describes the principles and steps of each algorithm in detail,and compares the advantages and disadvantages of each algorithm.In order to verify the applicable data set range of the algorithm,the algorithm is applied to different data set fields to prove the universality and efficiency of the algorithm.In view of the traditional KNN algorithm learning is the distribution of similar data around it,while the Bayesian method learning is the global distribution of data.Combining KNN with Bayesian method,that is,multi label k-nearest neighbor algorithm(ML-KNN),it not only inherits the advantages of the two,but also overcomes the impact caused by the imbalance of sample data.However,ML-KNN algorithm does not take into account: in the sample space,with the change of the distance from the nearest neighbor to the predicted sample point,the weight thatshould be allocated will also change.In order to solve this problem,an improved ML-KNN algorithm is proposed,which is named RML-KNN algorithm.The algorithm has been applied to the field of traditional Chinese medicine,and has achieved good results.It has proved that the algorithm model constructed from case data is reliable,and the process of diagnosing insomnia in traditional Chinese medicine has been pushed forward to an objective and scientific way.In view of the imbalance of the original insomnia data set,five syndrome types of insomnia were separated,namely 10 labels.After the separation,because the secondary label is too sparse to be learned by the model,the prediction results are not ideal.Therefore,we propose an algorithm called LRMI based on label combination.Firstly,the standard of multi label data set disequilibrium and the method of calculating category disequilibrium are defined in DEML.At the same time,a random strategy is used to build a balanced label subset.Then the multi label problem can be transformed into the multi classifier problem.Finally,all the multi classifiers are integrated to get the prediction results.