Computer Predicts Teenage Binge Drinkers

A computer may be able to do what a D.A.R.E. counselor can’t. A new study published in Nature shows scientists were able to teach a computer to predict which adolescents would become binge drinkers in the future with 70 percent accuracy. Using brain function, personality, life history, and genetics data from hundreds of 14-year-olds who were binge drinkers by the age of 16, the computer was able to “learn” 40 different variables associated with binge drinking. "It’s a form of supervised learning," says lead study author Robert Whelan. He said a computer may never be able to predict alcohol abuse with 100 percent accuracy, but his research can help focus on “catching them [teens at risk] so we can perform early interventions.”