Context-Recognition And Data Mining Methods For A Health Club Application

In this paper, a context-aware Health Club for cyclists is described. The Health Club is a tool for coaches or ordinary keep-fit enthusiasts to monitor the development of fitness and the improvement of tactics. This application is being built on Genie of the Net, a general software platform for context-aware applications. The work on context recognition was started by collecting data from cycling exercises. Data mining was used to find interesting and useful contexts from the collected data. Multilayer perceptrons and recursive least-squares were used in recognizing contexts. The contexts were utilized in analyzing the exercises. Different techniques for visualizing the exercise data, the recognized contexts and the results of the analysis were studied as well. A user interface was also built for testing the visualization techniques.