The idea behind Booost, is that upon registering for the app, an avatar for the user is created, which the user then cares for and talks to, developing a relationship. Over time, the app learns about the user and is able to analyze their conversations to determine one's mood and especially if there are signs of distress present. If so, the app then suggests ways to seek help. The hope of the app is that it leads to more timely diagnoses and treatments, and helps address the high rates of undiagnosed and untreated mental illness.
"The aim is to make it so Booost is “not just an app but a game, a friend,” says Nare Vardanyan, founder of Mindbin, who also describes the app as “a Siri with empathy… or a Tamagotchi with artificial intelligence.”
Claiming an 80% accuracy of successful emotional detection, Booost works through a mood and a semantic engine which allocate sentiment values to words- ‘pleased’ being a weak positive, and ‘grief’ being a strong negative, for instance. This in turn allows the app to analyse the conversations it has with the user and determine their mental state. If an issue is detected, the app would then suggest avenues of address such as encouraging therapy or providing numbers to helplines, or notify guardians in the case of young children."
For the full story click here: