The Future of Youth Tobacco Use: Predictive Clustering and Time Series Analysis

This research examines youth tobacco use in America today and in the future through the application of machine learning and statistics. This research applies linear regression and ARIMA techniques to predict future trends based on data collected by the CDC Youth Tobacco Survey. K-means clustering analysis also examines differences in youth tobacco use among states. This research concludes that youth tobacco use has been steadily decreasing over the years, from 1999 to 2017. Forecasts for 2018 to 2020 show three distinct groups of states.