ABSTRACT
As far as modelling is concerned, soccer is one of the most thrilling and challenging sports. The basic reason for the problem of modelling soccer is the data. This chapter summarizes the current state of art on challenges and subjects of modelling as it relates to sports. It introduces the concepts of modelling as well as methodological experience. The chapter presents the complexity of processes and tactical patterns as well as quantitative vs. qualitative information. It then introduces some approaches dealing with positional modelling of events, situations and processes in soccer. There are mainly two types of artificial neural networks in use: feed forward or backpropagation networks are used for multidimensional function calculation, while self-organizing maps (SOMs) are used for pattern recognition. Voronoi cells have been used for clustering on artificial neural networks of type SOM, where the cluster is the set of all neurons that, regarding the contained information, are most similar or closest to a representing prototype neuron.
