Software Designed to Detect Depression in Blogs

There are some who say that the way in which we communicate – as well as the words used – can say a lot about the way we feel at the time. For researchers at Ben-Gurion University of the Negev (BGU), depression can be detected in blogs and online texts.

This research was recently captured in a Science Daily release and showed that a software program can detect this depression. Developed by this team of researchers, this software is capable of identifying language that can indicate the psychological state of the writer and could serve as a screening tool.

The software was used to scan more than 300,000 English language blogs posted to mental health websites. The program then identified the 100 bloggers it perceived as the most depressed and the 100 bloggers that were the least depressed. The samples were then evaluated by a panel of four clinical psychologists who concluded there was a 78 percent correlation between the findings of the panel and the software.

According to Professor Yair Neuman, the program was designed to identify hidden depressive content that did not refer to obvious terms such as depression or suicide. A trained psychologist can spot various emotional states through intuition and now this program does this methodically with Web intelligence.

The software program is designed to detect words that express various emotions, such as colors that a writer will use to metaphorically describe certain situations. The word black, for instance, may be combined with another term to describe symptoms of depression, such as loneliness or sleep deprivation.

Neuman acknowledges that software cannot replace excellent human judgment, but it does provide a screening process to raise an individual’s awareness of his or her condition.