Abstract
Hypertension is one of the major diseases leading to as one of death in all over the world. In clinical practice, many epidemiological variables including environmental and physical factors could affect the disease. However, too many variables are not easy to collect, i.e. it is a time-consuming and expensive work. Recently, the relationship between anthropometric factors and metabolic syndrome is investigated more and more exact in the biomedicine studies. In this study, we will explore the significant anthropometric factors affected hypertension using data mining methods including Backpropagation Neural Network, Decision Tree and Rough Set. Three epidemiological criteria are evaluated for these methods. The results showed that six anthropometric factors are selected. They are waist circumference, right thigh circumference, left thigh circumference, volume of trunk, surface area of trunk and volume of right arm. Furthermore, the performance of decision tree is better than the other approaches. The result is not only useful for the physician as a tool for diagnosing hypertension, but it is sophisticated enough to be used in the practice of preventive medicine.