Kurzbeschreibung
In this paper the feasibility of distinguishing different activities - namely walking and doing Monty Python’s Silly Walks - using data collected with a smartphone is evaluated. Acceleration data of users performing both tasks is collected. The data is then processed and used to train i) a k-nearest neighbor classifier and ii) a long short-term memory neural network. We explain the framework we use for data collection and give an overview of both approaches regarding the implementation. Finally, we compare the performance of both approaches and state benefits and drawbacks of each method respectively.