Lucky Saleem, Student im Master Informatik der Universität Rostock, hat seine Masterarbeit zum Thema "Wearable Lifestyle Recommender Simulation and Test Environment" on 07.02.2023 erfolgreich verteidigt. Betreuer und Gutachter der Arbeit waren Angelina Schmidt und Prof. Michael Fellmann (beide von der Universität Rostock).
Kurzfassung:
A system that can manage in enhancing your daily lifestyle is needed because life is complex and in the professional world, there is rising complexity and flexibility of the user, making it very difficult to depend on all requirements. Wearable lifestyle recommender is a system that seeks to give context-sensitivity and personalization for people to better manage their lives, but because people are different, each one has different needs, making it impractical to offer each user individualized help. To test and set up personal or individualized recommender systems before using them, a personalization and test environment is required. A system that can adapt services to individual user requirements and preferences has the potential to improve service. Allowing users to modify the service themselves is one method to improve it; another is to proactively tailor services based on information provided by users or inferred from their prior behavior. These strategies work best when people are aware of what they want and need and when their preferences and actions are constant across contexts and time. But people don’t always know what they need or desire, and they frequently change their choices. Furthermore, people frequently struggle to express their preferences in the level of depth needed for personalization. Furthermore, individual desires and motivations are not the best way to set the personalized systems because they are not always realistic. Finally, the features of the system must adapt to runtime user requirements. For example, the user will start with the initial configuration, and then the system will adapt to user specifications. To cope with such a situation, we are prioritizing user preferences so that users can tailor the system according to their own choices and goals. The goal of this thesis is to create a user-friendly simulation environment where users can establish goals for leading healthy lifestyles by receiving personalized and context-sensitive recommendations. This system is made up of various components. First, a component to specify user settings, where the user may set up his or her goals and the best times to receive notifications based on the categories. Second, a component for simulating sensor data is also included so that the user can test the system without having to worry about really recording sensor data (e.g. 10000 steps per day). A third component is one for retrieving tips or notifications to confirm recommendations made based on the user’s settings and readings obtained from sensor data.