Application Research Of Decision Tree Method In Medical Diagnosis And Prediction

With the continuous improvement and enrichment of various data acquisition devices,a large amount of data is generated every day,and this phenomenon is no exception in the medical industry.In the medical industry,a large amount of case data,patient diagnosis information data,prescription information data issued by doctors,and the like are generated every day.At the same time,related technologies for information mining for data are also constantly developing.How to apply relevant technologies of data mining to the medical industry,and to discover some new knowledge and rules,and more and more attention has been received.Because heart disease is a serious threat to human health,and research on heart disease worldwide is constantly going on.In the computer field,The expert system is highly inspiring and can use the knowledge and experience accumulated by domain experts to make logical reasoning and result judgment on related issues,and it is also flexible.However,the expert system lacks self-learning,and supplement in the system depends on the knowledge and experience of domain experts.Using the decision tree algorithm,you can overcome some of the shortcomings of the expert system.First,the decision tree is very self-learning,and its model construction relies only on data.At the same time,the decision tree has the characteristics of process interpretability compared with other machine learning algorithms.According to the model,the relevant rules and knowledge can be easily summarized,which is of great significance for the mining of new knowledge.In this topic,three decision tree models are constructed using ID3,C4.5 and CART decision tree algorithms respectively.Designed to find models with higher performance and higher prediction accuracy in heart disease case data.At the same time,because the algorithms have their own advantages and disadvantages,in order to continuously improve the prediction accuracy of the model,this paper combines the models of the three algorithms,and introduces the concept of model contribution,using the estimated values of the three models.Weighted summation plus correction values are combined to form the final prediction.The built model was tested using test data and the results also verified that the improvement was relatively efficient.At the same time,according to the constructed model,the attribute characteristics closely related to heart disease can be excavated,and a simple model of heart disease prediction can be constructed to assist medical practitioners and patients in the diagnosis and prevention of heart disease.