Data Mining in Diabetes Diagnosis and Detection

Abstract

Diabetes is a disease worrying hundreds of millions of people around the world. In the USA, the population of diabetic patients is about 15.7 million (Breault et al., 2002). It is reported that the direct and indirect cost of diabetes in the USA is $132 billion (Diabetes Facts, 2004). Since there is no method that is able to eradicate diabetes, doctors are striving for ways to fight this doom. Researchers are trying to link the cause of diabetes with patients’ lifestyles, inheritance information, age, and so forth in order to get to the root of the problem. Due to the prevalence of a large number of responsible factors and the availability of historical data, data mining tools have been used to generate inference rules on the cause and effect of diabetes as well as to help in knowledge discovery in this area. The goal of this chapter is to explain the different steps involved in mining diabetes data and to show, using case studies, how data mining has been carried out for detection and diagnosis of diabetes in Hong Kong, USA, Poland, and Singapore.