Random tree algorithm to analyse the relation between type of traumatic dental injuries and its demographic and predisposing factors - A cross-sectional study.

Abstract

Traumatic dental injuries (TDIs) have become the public dental health problem worldwide in children and adolescents. These injuries are complex and multifactorial in aetiology. This study was done with the aim to analyse the association of 'type of TDI' with its demographic and various predisposing factors in children by an advanced statistical method of machine learning (ML) of artificial intelligence (AI). The present study's data were gathered by conducting the observational cross-sectional study among index age-groups 12 and 15 years children of randomly selected schools of different geographical regions. Structured interviews and dental examinations performed were done to record the variables of TDIs in self-constructed proforma. The gathered data were analysed by employing the random-tree model of machine learning algorithm of IBM SPSS Modeler version-18 software. Molar-relationship (2.5), age (1.75), sex (1.5) and geographical region/area (~1.5) were the most important predictors (factors) for the determination of type of dental injury as shown by the random tree model, whereas clinical factors like overjet (0.75), lip-competence (0.5) and overbite (0.5) showed lesser importance in the determination of type of TDIs. Demographic factors (age, sex and geographical region) and one clinical factor (molar-relation) were found as the stronger factors for determining the type of traumatic dental injury in children.