Using support vector regression for assessing human energy expenditure using a triaxial accelerometer and a barometer

Abstract

Physical inactivity is nowadays defined as the fourth leading risk factor for global mortality. These levels are rising worldwide with major aftereffects on the prevention of several diseases and the general health of the population. Energy expenditure (EE) is a very important parameter usually used as a dimension in physical activity assessment studies. However, the most accurate methods for the measurement of the EE are usually costly, obtrusive and most are limited by laboratory conditions. Recent technological advancements in the sensor technology along with the great progress made in algorithms have made accelerometers a powerful technique often used to assess everyday physical activity. This paper discusses the use of support vector regression (SVR) to predict EE by using a single measurement unit, equipped with a triaxial accelerometer and a barometer, attached to the subject's hip. © Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2013.