[Abstract]: In this paper, two novel distances for nominal time series are introduced. Both of them are based on features describing the serial dependence patterns between each pair of categories. The first dissimilarity ...