Categorical Time Series Analysis
Christian H. Weiß · Wiley StatsRef: Statistics Reference Online · 2018
Abstract Categorical time series are discrete‐valued time series having a qualitative range, which consists of a finite number of categories. As standard methods from real‐valued time series analysis cannot be applied to such a kind of data, tailor‐made tools for the visualization of the time series, characterizing the marginal distribution and measuring serial dependence are presented. Also, common stochastic models for such time series are surveyed, including types of Markov models, discrete ARMA models, Hidden‐Markov models, and regression models.