Hidden Markov model training

Rasmussen T.K. Krink, T. (2003). Improved Hidden Markov Model training for multiple sequence alignment by a particle swarm optimization-evolutionary algorithm hybrid. Bio Systems, Vol. 72, No. 1-2, pp. 5-17. 

The Hidden Markov Model (HMM) is a powerful statistical tool for modeling a sequence of data elements called the observation vectors. As such, extraction of patterns in time series data can be facilitated by a judicious selection and training of HMMs. In this section, a brief overview will be presented and the interested reader can find more details in numerous tutorials... 

The first step of the analysis is the training where Hidden Markov Models (HMMs) representing various operating behaviors are trained using labeled historical data from the process. In this section, three broad operat-... 

The regularization procedure takes a different form for each method of statistical learning. When training hidden Markov models, the derived probabilities do not only take the observed sequences into account but also use so-called prior distributions that formulate some hypothesis on the occurrence of output characters in the case that we have no additional information (such as the observed sequences). An expected background distribution of amino acids or nucleotides is a natural starting point for such a prior distribution. 

Woodland, P. C., and Povey, D. Large scale discriminative training of hidden Markov models for speech recognition. Computer Speech and Language 16 (2002), 25-47. 

The dynamic-system model is a natural choice for statistical generation of FO contours since it is well suited to the job of generating continuous trajectories. If it has any weaknesses, we can point to the facts that the state trajectories are limited to being those of a first-order filter, the noise terms have to be Gaussian and the traimng process can be quite intricate. An alternative is to use hidden Markov models (HMMs) since these are in general easier to train and allow more complexity with regard to noise/covariance terms. 

A trend analysis strategy is proposed that takes advantage of the wavelet-domain hidden Markov trees (HMTs) for constructing statistical models of wavelets (see Section 6.5). Figure 7.10 depicts the strategy that can be used to detect and classify faulty (abnormal) situations. As before, in the training phase, time series data collected under various conditions are... 

---