Medical data mining has been applied to accurate classification and rapid prediction for prognosis and diagnosis of patients in a specialized medical area (Masuda et al., 2002). It has been also used for training unspecialized doctors to solve a specific diagnostic problem (Kononenko, 2001). Among several algorithms for classification and prediction tasks, a decision tree is one of the most frequently used techniques in a medical data mining area. While it is easy to find many cases to prove the decision tree to be useful in the business domain, the decision tree enables to predict prognoses and diagnoses in the domain of medicine, using tree-structured models or in the form of ‘IF condition-based-on attribute-values THEN outcome-value’ to identify useful features of importance.
Yun (2008) utilized a C4.5 algorithm to build a decision tree in order to discover the critical causes of type II diabetes. She has learned about the illness regularity from diabetes data, and has generated a set of rules for diabetes diagnosis and prediction. Khan et al. (2009) used decision trees to extract clinical reasoning in the form of medical expert’s actions that are inherent in a large number of electronic medical records. The extracted data could be used to teach students of oral medicine a number of orderly processes for dealing with patients with different problems depending on time.