*2.2. Smartphone Applications*

Qian et al. [1] explored smartphone usage to predict sedentary behaviours. They were able to classify user contexts such as location, time and application usage to predict if the user would be sedentary in the coming hours. Their methodology is still unable to distinguish between different types of sedentary behaviour. Dantzig et al. [27] developed the SitCoach mobile application to monitor the physical activity and sedentary behaviour of office workers. The objective of their research was to avoid prolonged sitting by providing timely information to the user in terms of alert messages. They concluded that mobile applications can motivate people to take regular breaks from long sitting. Shin et al. [28] developed a mobile application to recognize user sedentary activity using a mobile device. Their method was based on rotated acceleration using quaternions, which classified sedentary behaviour with higher accuracy. However, their application required server-side processing to classify user activities patterns. It is seen that many systems and models were therefore proposed to track sedentary behaviour; however, they had limitations.

Our proposed approach is to expand upon the lessons learned from existing research work and to enable the detection of the contextual information by utilizing the embedded sensors of the smartphone and processing data in real time inside the smartphone environment.
