hmmlearn implements the Hidden Markov Models (HMMs). The HMM is a generative probabilistic model, in which a sequence of observable \(\mathbf{X}\) variables is generated by a sequence of internal hidden states \(\mathbf{Z}\). The hidden states are not be observed directly. The transitions between hidden states are assumed to have the form of a (first-order) Markov chain. They can be specified by the start probability vector \(\boldsymbol{\pi}\) and a transition probability matrix \(\mathbf{A}\). The emission probability of an observable can be any distribution with parameters \(\boldsymbol{\theta}\) conditioned on the current hidden state. The HMM is completely determined by \(\boldsymbol{\pi}\), \(\mathbf{A}\) and \(\boldsymbol{\theta}\).

There are three fundamental problems for HMMs:

  • Given the model parameters and observed data, estimate the optimal sequence of hidden states.
  • Given the model parameters and observed data, calculate the likelihood of the data.
  • Given just the observed data, estimate the model parameters.

The first and the second problem can be solved by the dynamic programming algorithms known as the Viterbi algorithm and the Forward-Backward algorithm, respectively. The last one can be solved by an iterative Expectation-Maximization (EM) algorithm, known as the Baum-Welch algorithm.


[Rabiner89]Lawrence R. Rabiner “A tutorial on hidden Markov models and selected applications in speech recognition”, Proceedings of the IEEE 77.2, pp. 257-286, 1989.
[Bilmes98]Jeff A. Bilmes, “A gentle tutorial of the EM algorithm and its application to parameter estimation for Gaussian mixture and hidden Markov models.”, 1998.

Available models

hmm.GaussianHMM Hidden Markov Model with Gaussian emissions.
hmm.GMMHMM Hidden Markov Model with Gaussian mixture emissions.
hmm.MultinomialHMM Hidden Markov Model with multinomial (discrete) emissions

Read on for details on how to implement an HMM with a custom emission probability.

Building HMM and generating samples

You can build an HMM instance by passing the parameters described above to the constructor. Then, you can generate samples from the HMM by calling sample.

>>> import numpy as np
>>> from hmmlearn import hmm
>>> np.random.seed(42)

>>> model = hmm.GaussianHMM(n_components=3, covariance_type="full")
>>> model.startprob_ = np.array([0.6, 0.3, 0.1])
>>> model.transmat_ = np.array([[0.7, 0.2, 0.1],
...                             [0.3, 0.5, 0.2],
...                             [0.3, 0.3, 0.4]])
>>> model.means_ = np.array([[0.0, 0.0], [3.0, -3.0], [5.0, 10.0]])
>>> model.covars_ = np.tile(np.identity(2), (3, 1, 1))
>>> X, Z = model.sample(100)

The transition probability matrix need not to be ergodic. For instance, a left-right HMM can be defined as follows:

>>> lr = hmm.GaussianHMM(n_components=3, covariance_type="diag",
...                      init_params="cm", params="cmt")
>>> lr.startprob_ = np.array([1.0, 0.0, 0.0])
>>> lr.transmat_ = np.array([[0.5, 0.5, 0.0],
...                          [0.0, 0.5, 0.5],
...                          [0.0, 0.0, 1.0]])

If any of the required parameters are missing, sample will raise an exception:

>>> hmm.GaussianHMM(n_components=3)
>>> X, Z = model.sample(100)
Traceback (most recent call last):
sklearn.utils.validation.NotFittedError: This GaussianHMM instance is not fitted yet. Call 'fit' with appropriate arguments before using this method.

Fixing parameters

Each HMM parameter has a character code which can be used to customize its initialization and estimation. EM algorithm needs a starting point to proceed, thus prior to training each parameter is assigned a value either random or computed from the data. It is possible to hook into this process and provide a starting point explicitly. To do so

  1. ensure that the character code for the parameter is missing from init_params and then
  2. set the parameter to the desired value.

For example, consider an HMM with explicitly initialized transition probability matrix

>>> model = hmm.GaussianHMM(n_components=3, n_iter=100, init_params="mcs")
>>> model.transmat_ = np.array([[0.7, 0.2, 0.1],
...                             [0.3, 0.5, 0.2],
...                             [0.3, 0.3, 0.4]])

A similar trick applies to parameter estimation. If you want to fix some parameter at a specific value, remove the corresponding character from params and set the parameter value before training.


Training HMM parameters and inferring the hidden states

You can train an HMM by calling the fit method. The input is a matrix of concatenated sequences of observations (aka samples) along with the lengths of the sequences (see Working with multiple sequences).

Note, since the EM algorithm is a gradient-based optimization method, it will generally get stuck in local optima. You should in general try to run fit with various initializations and select the highest scored model.

The score of the model can be calculated by the score method.

The inferred optimal hidden states can be obtained by calling predict method. The predict method can be specified with decoder algorithm. Currently the Viterbi algorithm ("viterbi"), and maximum a posteriori estimation ("map") are supported.

This time, the input is a single sequence of observed values. Note, the states in remodel will have a different order than those in the generating model.

>>> remodel = hmm.GaussianHMM(n_components=3, covariance_type="full", n_iter=100)
>>> remodel.fit(X)  
>>> Z2 = remodel.predict(X)

Monitoring convergence

The number of EM algorithm iteration is upper bounded by the n_iter parameter. The training proceeds until n_iter steps were performed or the change in score is lower than the specified threshold tol. Note, that depending on the data EM algorithm may or may not achieve convergence in the given number of steps.

You can use the monitor_ attribute to diagnose convergence:

>>> remodel.monitor_  
          iter=12, n_iter=100, tol=0.01, verbose=False)
>>> remodel.monitor_.converged

Working with multiple sequences

All of the examples so far were using a single sequence of observations. The input format in the case of multiple sequences is a bit involved and is best understood by example.

Consider two 1D sequences:

>>> X1 = [[0.5], [1.0], [-1.0], [0.42], [0.24]]
>>> X2 = [[2.4], [4.2], [0.5], [-0.24]]

To pass both sequences to fit or predict first concatenate them into a single array and then compute an array of sequence lengths:

>>> X = np.concatenate([X1, X2])
>>> lengths = [len(X1), len(X2)]

Finally just call the desired method with X and lengths:

>>> hmm.GaussianHMM(n_components=3).fit(X, lengths)  
GaussianHMM(algorithm='viterbi', ...

Saving and loading HMM

After traning an HMM can be easily persisted for future use with the standard pickle module or its more efficient replacement in the joblib package:

>>> from sklearn.externals import joblib
>>> joblib.dump(remodel, "filename.pkl")
>>> joblib.load("filename.pkl")  

Implementing HMMs with custom emission probabilities

If you want to implement other emission probability (e.g. Poisson), you have to subclass _BaseHMM and override the following methods

base._BaseHMM._init(X, lengths) Initializes model parameters prior to fitting.
base._BaseHMM._check() Validates model parameters prior to fitting.
base._BaseHMM._generate_sample_from_state(state) Generates a random sample from a given component.
base._BaseHMM._compute_log_likelihood(X) Computes per-component log probability under the model.
base._BaseHMM._initialize_sufficient_statistics() Initializes sufficient statistics required for M-step.
base._BaseHMM._accumulate_sufficient_statistics(…) Updates sufficient statistics from a given sample.
base._BaseHMM._do_mstep(stats) Performs the M-step of EM algorithm.