Researchers teach a computer to predict teen binge drinkers
Intervention during someone’s teenage years is frequently the key to preventing alcohol abuse in adult life. It’s good to know, then, that a group of scientists has found a way to predict that abuse at an early age using computer modeling. The approach teaches the compute how to spot a likely teen binge drinker by weighing 40-plus biological and social factors that include brain structure, any enabling genes, past events and personality traits. If a 13-year-old is already smoking because of an addictive personality or family influences, for instance, it’s more likely that this child will pick up a dangerous drinking habit a few years later.
The early technology produced some false positives, but it was good enough to predict the likelihood of binging with 70 percent accuracy. It could become more reliable if given time — lead researcher Robert Whelan tells The Verge that he’d like future modelling to account for peer pressure from social networks, which wasn’t a major concern when prepping the study years earlier. While it’s doubtful that computer predictions will ever be completely accurate, they might get close enough that concerned parents and schools would often know when to take action.
[Image credit: Spencer Platt via Getty Images]
Filed under: Science
Via: The Verge