RDAHMM

From Quakewiki

RDAHMM is a softare implementation of the regularized deterministic annealing expectation-maximization (RDAEM) algorithm for fitting hidden Markov models (HMMs) [1]; fitting an HMM to a time series allows us to describe the statistics of the data in a simple way that ascribes discrete modes of behavior to the system. Unlike standard HMM time series fitting methods, such as those used in speech analysis and synthesis, the RDAEM approach does not require a priori knowledge about the data to provide high-quality, self-consistent model fits (although it does not exclude the use of such information where available). This means that it can be applied to data collected from systems that are poorly understood, or when avoiding solution bias is a high priority. In addition, it means that the approach can be quickly adapted to new problem domains.