New Smartphone Software Predicts Users' Mood

Aug. 7, 2013— -- The smartphone could end up moonlighting as a mood ring in addition to its job as a computer. MoodSense, a new software package in development by Microsoft Research and Rice University in Texas, doesn't rely on a camera to capture your smile or a microphone to hear you scream in anger. Instead, it predicts moods from what you do with your phone.

Lin Zhong, an associate professor of electrical and computer engineering at Rice, saw mood as an important factor in how people interact with one another. "For example, when we deal with a spouse, it's very important to gauge [his or her] mood to make sure you have a smooth relationship," he told ABC News.

But asking someone about their mood or even saying how you feel out loud could be considered intrusive and even rude. "It's become a little obnoxious to tell people [on Facebook or Skype] whether you're feeling great or sad," said Zhong. "If a computer can do it automatically, it's a more socially acceptable way to share your mood."

Zhong and graduate student Robert LiKamWa developed MoodSense to track how smartphone users spend their time across different activities such as email, websites and phone calls. Every couple of hours, the software asks users to rate both their happiness and their activity level on a scale of 1 to 5. Eventually, MoodSense gathers enough data to predict how users feel according to their recent activity. After the orientation period, the researchers said MoodSense could predict a person's mood with 93 percent accuracy.

Shiny Happy People Have a Site to Call Their Own

Md. Munirul Haque, a computer science postdoctoral fellow at the University of Alabama-Birmingham, also developed mood prediction software, called IRENE. IRENE, however, used either a phone camera or a webcam to take a snapshot and predict what mood users were in, which he acknowledged had its problems. "People can easily fake their moods in their pictures," he wrote in an email. "But it is highly improbable to fake mood through hundreds of activities being done in a smartphone."

Zhong sees the software package as a way to avoid awkward moments in digital conversations. "When someone is calling, they'll know the person's mood beforehand and be able to better handle that call," he said. In addition, MoodSense will also create an API that will broadcast moods the same way GPS devices broadcast location. App developers can use the API and incorporate mood-based functions into their software.

Ultimately, it's up to the smartphone developers to decide whether or not they want MoodSense as part of their software package. "I hope it will be available in a couple of years," said Zhong. "But I'm only a professor. I'm not the VP of Microsoft."