The aim of the project is to access the feasability of using devices worn on wrists, such as smartbands and wristwatches, to predict whether the user is prone to depression. Our submission hopes only to aid in the early detection of depression so that the affected people can seek timely aid.
We use a variety of methods such as rocket transformers for convolution, classification on series using RandomForrests, BOSS and Ridge regression to aid us in our core questions about the relation of activity and depression.