Exploration of risk factors associated with adolescent alcohol consumption using cutting edge recursive partitioning techniques

The purpose of this article is to explore and identify risk factors influencing drug use in school going adolescents aged 10 to 19 in a hilly state in the North-Eastern part of India. This article will explore the data collected from the National Institute of Health and Family Welfare, New Delhi, by using cutting edge Recursive Partitioning techniques such as Discriminant Analysis, Decision Tree Method, Artificial Neural Network and the Stochastic Gradient Boosting to build a predictive model. Out of 3069 randomly selected participants who undertook the Adolescent Reproductive and Sexual health (ARSH) questionnaire a subset have been used to form this data set. Utilization of Artificial Neural Network, Stochastic Gradient Boosting and the Random Forest models produce higher accuracy and classification in contrast to other measures. These models will be useful in the prediction of associated risk factors that contribute to adolescent alcohol consumption.