Research And Application Of Optimized Fuzzy Decision Tree In The Classification Of Medical Data

At present,with the rapid development of information technology and the continuous popularization of information technology achievements,many hospitals have been generally equipped with information systems and digital equipment.With the increase of the doctor,the relevant data about the database the patients also constantly accumulated,such as image data(such as CT image data),signal(ecg),digital data(such as blood,urine test results),text data(as described in the doctor diagnosis results,patients with symptoms)and so on the different types of medical data.Due to different levels of doctors or inconsistent subjective judgments,misdiagnosis may occur in the process of diagnosis.Therefore,it is more and more urgent for medical data mining to obtain the contained information.By mining these medical data,valuable association rule information can be obtained,which can provide more scientific basis for doctors to diagnose diseases and reduce the rate of misdiagnosis.Therefore,it is of great research significance and application value to classify medical data and mine association rules.The final method to classify medical data is to choose a good classification model.The traditional data classification usually USES the decision tree model as the classification model,but the decision tree model cannot apply to the fuzziness and uncertainty in the data.Therefore,the construction of a fuzzy decision tree model by combining the fuzzy theory with the decision tree model can greatly enhance the applicability of the decision tree model to the ambiguity and uncertainty in the processing of medical data,so as to classify and mine medical data more accurately and effectively.Firstly,two kinds of decision tree algorithms,fuzzy decision tree algorithm and decision tree algorithm,are compared.The construction and rule extraction of fuzzy decision tree and decision tree were completed on the medical data set,and the differences between fuzzy decision tree and decision tree were summarized.Kohonen feature mapping algorithm was used to determine the central point of the medical data set,and the distance between each data and the central point was calculated.and the processed data set was constructed by the fuzzy decision tree model.By comparing with other decision tree models in classification accuracy and the number of rules generated,the superiority of fuzzy decision tree in medical data classification is proved.Secondly,the particle swarm optimization algorithm is improved by analyzing and comparing the linear particle swarm optimization algorithm with fixed weight.In order to set the inertial weight parameters,a nonlinear descending particle swarm optimization algorithm is proposed.Four kinds of standard functions are used to test the algorithm,which can better jump out of the local optimum in the early iteration and conduct global optimization,and better local optimization in the late iteration to speed up the convergenceBy combining the improved particle swarm optimization algorithm and fuzzy decision tree algorithm in this paper,a hybrid method for medical data set classification,the method using particle swarm optimization algorithm to optimize the key parameters of fuzzy decision tree,construct an optimal parameters in fuzzy decision tree,and applied in the data set on the classification of breast cancer,mining the association rules.In order to prove the effectiveness of the model proposed in this paper,Diabetes data set and Breast Cancer data set on the UCI public medical data set are used as experimental data sets to optimize the fuzzy decision tree model.Through experiments,the comparison of training accuracy,test accuracy,generalization ability and spanning tree scale composition shows that the model proposed in this paper has a good performance in medical data classification.