The analysis of nominal data is often reduced to accumulation and description. Bayesian methods offer a possibility to analyse nominal data in a more sophisticated way. The possibility to indicate a structure via graphical representation, where variables are nodes and relationships are edges, enriches this method and makes it a powerful tool for data analysis. In this paper, an overview on Bayesian methods is given, the underlying rule is presented and some specialities will be discussed. Bayesian belief networks are described in brief and their potential to use them in case of uncertainty is presented. This includes not only the methods, but also possible applications in this context.