Data Mining Based Evaluation Method For Hypertension Disease

The control of hypertension is a big worldwide problem for public health. The goal for the treatment of hypertension is to decrease the blood pressure, but the aim is to reduce the total risk of cardiovascular disease. In clinic, the control of hypertension is not very good. The key reason for this problem is the lack of an effective method to evaluate the cardiovascular status of hypertensive.This study established an evaluation model for cardiovascular status based on data mining. The inputs are physiological parameters from clinic and the outputs are eigenvalues for evaluating the status of cardiovascular system. Based on this model, a method for evaluating hypertension disease was founded. Accordingly, the following tasks were completed:1. The cardiovascular parameters were analyzed by clustering analysis.2. The peripheral circulation evaluation index was calculated from the peripheral circulation parameters by artifical neural network.3. The cardiac function evaluation index was obtained by adding the cardiac function parameters with weighted value.4. The above two indexes were utilized to establish a parameter plane. The plane was divided into 9 areas to indicate various cardiovascular statuses by 'Youden index'.5. The fuzzy logic was used to describe 3 contours of fuzzy membership on parameter plane. Also fuzzy membership of various cardiovascular statuses could be achieved.The cardiovascular parameters were divided into peripheral circulation parameters and cardiac function parameters successfully by clustering analysis. The resultobtained by testing 342 subjects showed that the cardiovascular status of norm blood pressure, prehypertensive and hypertensive subjects had distinct distribution on the parameter plane. The analysis of individual could work out the cardiovascular risk stratification, the peripheral circulation evaluation index, the cardiac function evaluation index, the corresponding area from 'Youden index' and the fuzzy membership of cardiovascular status. Then these above could evaluate the cardiovascular status objectively.In conclusion, the model offers an original idea for evaluation of hypertension disease and could assist diagnosis in clinic for hypertension and other chronic diseases after improvement.