Analysis of Smoking in Korean Adolescents Based on Data Mining Techniques

Abstract

The purpose of this study was to analyze the risk factors of adolescent smoking using data mining techniques. Data were from the 2012 Korea Youth Risk Behavior Web-based Survey. Subjects were 36,889 high school students (18,938 men, 17,951 women). A prediction model was developed by the use of a Classification and Regression Tree Algorithm. In the decision tree analysis, academic performance, suicidal ideation, type of school they attend and indirect smoking experience were significantly associated with adolescent smoking.