In order to give the reader an idea of the versatility of HMMs, some of the areas of application of HMMs are listed in Table 14.1. This list is by no means exhaustive; in fact, several hundred papers on applications of HMMs had already been published by 2001 (Cappé, 2001). It is worth noting that, in some of these applications, the states of the HMM have a clear interpretation (e.g. for capture-recapture data, in which case the state indicates whether or not an animal is alive); see Chapter 24. In the context of supervised learning as discussed in Section 5.6, the states are defined a priori and observed in a training sample.