Researchers from Harvard Medical School have used machine learning to predict the 5 percent of U.S. Army soldiers who later committed one third of all violent crimes in the workplace between 2004 and 2009.
Their analysis of data from 2011 to 2013 was even more accurate, predicting the 5 percent of soldiers who would commit 50.5 percent of violent crimes.
The researchers say updating and cleaning the more than 3 terabytes of data on military personnel from many different datasets was a major part of the project. The machine-learning algorithms applied to the data depended on the different risk factors for soldiers and variable changes for each individual and over time.
More than 50 algorithms were available and none of them were appropriate for every application, "so the issue is selecting the right one for an application," says Harvard professor Ronald Kessler.
He reports the team is now looking to sharpen its understanding of causal links, and hopes to focus on the appropriate treatments for depression in the future.
From Government Computer News
View Full Article
Abstracts Copyright © 2015 Information Inc., Bethesda, Maryland, USA