Background: Sedentary behavior has been associated to the development of noncommunicable diseases (NCD) such as cardiovascular diseases (CVD), type 2 diabetes, and cancer. Accelerometers and inclinometers have been used to estimate sedentary behaviors, however a major limitation is that these devices do not provide contextual information such as the activity performed, e.g., TV viewing, sitting at work, driving, etc.
Objective: The main objective of the thesis is to propose and evaluate a Personal Health Record System to support the assessment and monitoring of sedentary behaviors.
Results: Until now, we have implemented a system, which identifies individual's sedentary behaviors and location based on accelerometer data obtained from a smartwatch, and symbolic location data obtained from Bluetooth beacons. The system infers sedentary behaviors by means of a supervised Machine Learning Classifier. The precision in the classification of the six studied sedentary behaviors exceeded 90%, being the Random Forest algorithm the most precise.
Conclusion: The proposed system allows the recognition of specific sedentary behaviors and their location with very high precision.