In partnership with the University of Exeter, we are pleased to announce that Activinsights will be sponsoring a PhD studentship specialising in raw data research.
The overall aim of this project is to use a combination of mathematics, statistical analysis and software development to automatically identify the activity (sitting, walking, running, etc) that the wearer of a tri-axial accelerometer is engaged in. Features that characterise the activity will be derived both theoretically (using knowledge about the position of the accelerometer) and by machine learning of “sparse features” from the large amounts of data available. With the features on hand, we plan to use models such as hidden Markov models, switching state-space models and conditional random fields to map the temporal patterns of features into activities. You will have the opportunity to use and develop several state-of-the-art machine learning techniques applied to “big data”, whilst also using a variety of programming languages (such as R, MATLAB and Python). The main output from the project will be an open source software program to automatically segment accelerometer data into clinically meaningful physical activity classifications according to type, frequency, duration and intensity that can be used by researchers and clinicians.
If you know of anyone who may be interested in this fully funded PhD position please respond directly to the advert below before the end of the month.