查看文章

Our driving vision is a smartphone service, called Mood-Sense, that can infer its owner’s mood based on information already available in today’s smartphones. The service will fundamentally enhance context-awareness by providing clues about mobile users’ mental states. We report early results from studying 25 iPhone users in the field and the correlation between their mood and phone usage. We show that user mood can be inferred into four major types with an average accuracy of 91%. This is achieved using only three weeks of training data and simple smartphone usage statistics. The results, though preliminary, strongly suggest the feasibility of mood inference without using the powerhungry and socially invasive microphone and camera.