Home / News, Videos & Publications / News / Homeland & Cyber Security /

Using Science to Prevent School Shootings

Using Science to Prevent School Shootings

September 16, 2015

Homeland & Cyber Security

Forbes — The rash of school shootings in the last decade, from Virginia Tech in 2007 to Sandy Hook in 2012, has spawned an ongoing debate about whether troubled students should be found and dealt with before their anger turns them into school shooters.

Now a team of researchers is proposing a controversial solution: They want to use science to analyze students’ blog posts — in essence using words to separate those who are merely angry from those who are truly dangerous.

A new technology called “natural language processing” (NLP) can identify patterns of words that the researchers believe would signal impending violent behavior.

Prof. Yair Neuman

“We’re excavating the deeper layers of text to determine which people are expressing the highest levels of depression” indicative of potential violent behavior, says lead author Yair Neuman, a professor in the Homeland Security Institute at Ben-Gurion University.

The key to making NLP work in the context of school safety is to plot key words in blog posts against neighboring words, and then look for patterns that indicate violent tendencies, Neuman explains. For example, the word “depressed” would only be suspicious if it appeared in close proximity to words such as “lonely,” “anxious” and “suicidal.”

Neuman acknowledges that implementing NLP in schools would present both ethical and practical challenges, but contends that NLP technology could pick up on clues that parents might otherwise miss.

“High school can be a very depressing place. Think about the kids being bullied or victimized by humiliation,” he says. “Parents are not experts. They don’t have access to the full emotions of their teenagers.”

Read the full article on the Forbes website >>