Title: Prediction of underage alcohol, marijuana, and inhalant use as predictors
Authors: Sarah Best; Jong-Min Kim
Addresses: Statistics Discipline, Division of Science and Mathematics, University of Minnesota-Morris, Morris, Minnesota, USA ' Statistics Discipline, Division of Science and Mathematics, University of Minnesota-Morris, Morris, Minnesota, USA
Abstract: Since the 1960s, drug use among American teenagers and young adults has been a growing, public concern. So we investigated how proportions of underage alcohol, marijuana and inhalant users differ by age and substance type, the manner in which underage alcohol, marijuana and inhalant use is distributed across age groups, and how age and use of other substances impact the odds of an underage individual having used a substance. Confidence intervals of their true population proportions were calculated from the survey results of 70,000 Americans. Logistic regression models were then constructed and analysed to determine the relationship between age and substance use, and the predictive ability of the logistic regression model was compared to that of random forest and K-nearest neighbour models.
Keywords: drug use; logistic regression; K-nearest neighbours algorithm; random forest.
DOI: 10.1504/IJBHR.2024.144274
International Journal of Behavioural and Healthcare Research, 2024 Vol.9 No.3/4, pp.202 - 218
Accepted: 22 Aug 2024
Published online: 04 Feb 2025 *