A method for training and applying a machine learning algorithm to assess and predict physical health of a human user based on measurement of physical motion of the human user, the method comprising:
- assessing and categorizing a physical health state of a sample user;
- providing a first motion capture system comprising a force plate, a virtual reality headset, and two handheld virtual reality controllers;
- instructing the sample user to stand on the force plate, wear the virtual reality headset, and grip the two handheld virtual reality controllers;
- instructing the sample user to stand as still as the sample user is able to, for a plurality of short durations of time, while the motion capture system records a set of sample user motion data regarding slight bodily motion of the sample user even while the sample user is attempting to stand still;
- filtering and batch processing the set of sample user motion data;
- training a machine learning algorithm to predict a sway pattern of the sample user based on the set of sample user motion data;
- training a machine learning algorithm to associate the sway pattern of the set of sample user motion data as characteristic for the physical health state of the sample user as previously assessed;
- exporting the machine learning algorithm as trained to a virtual reality application;
- and
- applying the machine learning algorithm as trained to predict an unknown physical health state of a new user based on assessment of a new user sway pattern in a received motion data set associated with the new user.