Implementing spectral methods for hidden Markov models with real-valued emissions

@article{Mattfeld2014ImplementingSM,
  title={Implementing spectral methods for hidden Markov models with real-valued emissions},
  author={Carl Mattfeld},
  journal={ArXiv},
  year={2014},
  volume={abs/1404.7472}
}
Hidden Markov models (HMMs) are widely used statistical models for modeling sequential data. The parameter estimation for HMMs from time series data is an important learning problem. The predominant methods for parameter estimation are based on local search heuristics, most notably the expectation-maximization (EM) algorithm. These methods are prone to local optima and oftentimes suffer from high computational and sample complexity. Recent years saw the emergence of spectral methods for the… CONTINUE READING

References

Publications referenced by this paper.
SHOWING 1-10 OF 13 REFERENCES

vol

Sajid M. Siddiqi, Byron Boots, Geoffrey J. Gordon. Reduced-Rank Hidden Markov Models. arXi preprint
  • arXiv:0910, October
  • 2009
VIEW 26 EXCERPTS
HIGHLY INFLUENTIAL

Proceedings of the thirty-seventh annual ACM symposium on Theory of computing STOC ’05

Elchanan Mossel, Sébastien Roch. Learning nonsingular phylogenies, hidden Markov models
  • page 366,
  • 2005
VIEW 3 EXCERPTS
HIGHLY INFLUENTIAL

vol

Richard A Redner, Homer F Walker. Mixture Densities, Maximum Likelihood, the EM Algorithm. SIAM Review
  • 26, no. 2, pages 195–239, April
  • 1984
VIEW 3 EXCERPTS
HIGHLY INFLUENTIAL

vol

Daniel Hsu, Sham M. Kakade, Tong Zhang. A spectral algorithm for learning Hidden Ma Computer, System Sciences
  • 78, no. 5, pages 1460–1480, September
  • 2012

vol

Herbert Jaeger. Observable operator models for discrete st computation
  • 12, no. 6, pages 1371–98, June
  • 2000

MA

Christopher D Manning, Hinrich Schütze. Foundations of statistical natural langua Press, Cambridge
  • USA,
  • 1999
VIEW 2 EXCERPTS