Research And Aplication On Lung Diseases Cost Analysis Of Comentropy Based Decision Tree Algorithm

The analysis of developed decision tree is the process of knowledge discovery. Performance of decision tree depends on model tree's complexity and predicting accuracy. Decision tree is generated according to heuristic rules, the common decision tree generation algorithm is based on information theory ID3, C4.5 algorithms, but these algorithms have shortcomings in the variety of practical application. This article presents a discretization algorithm based on information entropy I-C4.5 algorithm to reduce large computational problems of data processing in C4.5 algorithm.This article describes the application of improved algorithm I-C4.5 in hospital cost data on lung diseases in data mining. And comparing analysis of the I-C4.5 algorithm and the C4.5 decision tree algorithm, after that, it generates rules to verify the I-C4.5 decision tree algorithm in the superiority of rule mining. The probability of hospitalization of patients through the lung response to its demand for medical services, hospital costs in patients with lung disease patients in the hospital reflects the health of their utilization, and then tap the probability of pulmonary disease and hospital inpatient costs of factors to help understanding of medical practice in pulmonary disease better and pulmonary disease patients in hospitals to improve service delivery basis.