Extracting Scenario for Diabetes Prevention from Daily Life Information

Abstract

It is important to understand scenarios deciding if you catch a disease or not. This paper supposes a method to extract such important scenarios by combining Decision Tree and KeyGraph visualization. Decision Tree is used to find effective attributes dividing dataset into clusters towards objective variable. KeyGraph is used to find important rare events between the clusters. Dataset is focused by data found by DT and KG step by step and after a few iterations meaningful events would be visualized. This method is applied for Health and Retirement Study data to extract not usual but significant scenarios for diabetes prevention.