Rule extraction of cardiovascular database using decision trees

Abstract

This study analyses the rule extraction of the cardiovascular database using classification algorithm. Heart disease is one of the most common causes of death in the western world and increasing so in Cyprus. We constructed a user-friendly data mining application based on the decision tree classification algorithm, so rules can be extracted from the decision tree models. Building the decision tree various splitting criteria methods were used. Using the database, we compared the performance of each splitting criteria method. The extraction of rules is done from the decision tree model with the highest performance. Our goal is to try and reduce this high number of casualties, by identifying the main causes of heart attack.