Russian artificial brain will predict the probability of death by indications of fitness-trackers
The basis of the new method is the estimation of physical activity of testees. The authors of the study have managed to connect some definite patterns of movements with particular variations of health state. They have used data from National Health and Nutrition Examinations Survey (NHANES) collected in the USA in 2003−2004 and 2005−2006.
The authors of the work have compared the level of physical activity of probationers of NHANES and patterns of their movements with the results of medical tests of the same people and, if this number is available, with the lifetime. They have found a connection between these groups of data using deep machine learning. Based on this AB, a mobile application has been already created where a user can independently analyze possible risks.