Recommendation system for human physical activities using smartphones

Important information can be obtained from smartphone users data such as profile modeling, behavior recognition, geolocalization, etc. Human activity recognition (HAR) from sensor smartphone data is a field which has garnered a lot of attention due to its high application in various domains such as the user health. In this paper, we will consider data from accelerometer to recognize the kind of user movements that we will classify to six kinds using machine and deep learning algorithms. Then, based on these results, we will make a recommendation system to inform the users of smartphone about their healthy behavior related to their physical activities.