Modules#

Summary of modules of nomopy

  • Factorial hidden Markov models (nomopy.fhmm):
    • FHMM: Implementation of a factorial hidden Markov model, including (1) several expectation-maximization (EM) algorithms for model parameter fitting, (2) confidence interval generation, (3) hidden state trajectory reconstruction, and (4) data generation

    • FHMMCV: Machinery for cross-validation and model selection through train-test splitting of noise timeseries

    • Hessian: Machinery for computing the Hessian of the loglikelihood with respect to model parameters

  • Higher order statistics (nomopy.hos):
    • second_spectrum(): Evaluation of the second spectrum, as defined in [SeidlerSolin1996]

    • chi2_test_gaussianity(): chi-squared test for Gaussianity of the noise timeseries, based on the second spectrum

  • Noise models (nomopy.noise):
    • ThermalTLFModel: Model for an ensemble of thermally-activated two-level fluctuators

[SeidlerSolin1996]

G.T. Seidler and S.A. Solin, Physical Review B 53, 9753 (1996)

nomopy.fhmm.FHMM#

Implementation of a factorial hidden Markov model, including (1) several expectation-maximization (EM) algorithms for model parameter fitting, (2) confidence interval generation, (3) hidden state trajectory reconstruction, and (4) data generation

nomopy.fhmm.FHMMCV#

Machinery for cross-validation and model selection through train-test splitting of noise timeseries

nomopy.fhmm.Hessian#

nomopy.hos#

nomopy.noise#