Classification Tree Method for Determining Factors Associated with Halitosis

Decision trees are data mining techniques for extracting hidden knowledge from large databases. This study was performed to establish the risk factors associated with halitosis by applying a decision tree model in Turkish population and examining interactions between these factors. We obtained total of 1.290 patients, consisting 645 patients healthy controls. The subjects’ demographic characteristics, smoking status, alcohol intake, medical history medications were assessed. presence potential intraoral causes determined investigating perceived oral health problems such as caries, periodontal diseases, tongue coating, cavity pathologies. Halitosis level evaluated using an organoleptic scale. All subjected classification analyses. significantly more common (80.9%) than without (20.7%) (P < 0.001). less non-smokers smokers (14.5%; P .001). evident all problems, smokers, those respiratory diseases (100%). effects systemic on significant 0.05). Respiratory conditions showed 0.01). developed identify halitosis. method that most affecting followed status.