Research On Data Mining Method Of Medical Record In TCM Based On Integrated Learning

With the development of computer science technology and information technology,TCM is also gradually following the trend to achieve information and modernization.And the dialectical classification technology of TCM syndromes and syndromes has attracted wide attention and made corresponding development.It is one of the main research topics in the field of traditional Chinese medicine.Due to the complexity,ambiguity and uncertainty of TCM case data,traditional single classification mining cannot ensure comprehensive consideration of all information.Considering that ensemble learning has a higher classification accuracy and better generalization performance,the ensemble learning model is used to classify the data of asthma cases provided by cooperative hospitals to obtain the dialectical rules of TCM asthma data.First of all,the TCM medical record was quantified,and the main symptoms of TCM asthma were obtained according to the improved feature selection algorithm IHCFS(Improved Hierarchical Clustering Feature Selection Algorithm).In order to describe the symptoms of medical records in Chinese medicine more objectively,the symptom information of medical records is graded and quantified,and the digital processing of automatic batch text is realized by database programming for subsequent research.For the characteristics of traditional Chinese medical record data features,there are many symptoms and redundant information,etc.The IHCFS algorithm is improved to get the main symptoms of asthma by improving the evaluation function and the termination criterion of the hierarchical clustering feature selection algorithm.Through simulation experiments,it is proved that the acquired main symptoms are more conducive to subsequent dialectical classification.Then,an ensemble learning based on Sampling and Future Selection(ELSFS),which is an ensemble learning based on multi-modal perturbation strategy,is proposed for the dialectical classification of TCM medical records.The random sampling method is used to perform sample spatial disturbance,and the IHCFS algorithm is used to perform feature space perturbation to get a larger base classifier.Simulation results on the UCIdataset show that the ELSFS algorithm has better classification performance.Finally,considering the excessive number of base classifiers will produce redundant base classifier to affect the prediction performance of the model,a Selective Ensemble Based on Weighted Greedy Strategy(SELGS)is proposed to reduce the storage space in the prediction stage,so as to improve the classification ability.The simulation shows that the SELGS algorithm has better dialectical classification performance.