Research On Medical Data Visualization Based On Feature Selection And Decision Tree Algorithm

With the explosive growth of data,visual medical treatment based on medical data is also developing rapidly.It becomes increasingly crucial to do research on it.Through the research on medical data visualization and related data classification algorithms,this paper proposes the design of decision tree model based on feature selection and the improved model of decision tree based on visual statistics.The specific research work is as follows:Ⅰ.From the design of decision tree model based on feature selection,we made a research on ID3 and Relief feature selection algorithm of decision tree.Since the decision tree ID3 algorithm is easy to assign attributes of low importance to high information gains,and the Relief feature selection algorithm can filter out attributes of low importance,the model is constructed with the advantages of the two algorithms combined.Through the decision tree ID3 algorithm classification,the first result attribute is obtained,and the result attribute is filtered for the second decision tree ID3 algorithm.Then the Relief feature selection algorithm is used for attribute screening to supplement the decision tree ID3 algorithm.The irrelevant attributes are screened out without increasing the complexity of the algorithm in the following,and the accuracy of the decision tree algorithm is improved through dimensionality reduction.Ⅱ.The improved decision tree model based on visual statistics.Using visual statistics to process medical data to obtain relevant important attributes,combined with the decision tree model of feature selection for comparative analysis of before and after results attributes.Screening and supplementing the relevant and potentially irrelevant attributes that may be screened out.After screening and supplementing,the decision tree ID3 algorithm is classified again with more comprehensive important attributes.The accuracy of the model when making decisions through the experiment of decision tree ID3 algorithm.