predict_mhmm()

predict_mhmm() assigns each sequence to its most likely mixture cluster (hard classification).

Function Usage

predict_mhmm(
    model,
    newdata=None,
    compress=False
)

seqHMM Parameter Mapping

Sequenzo	seqHMM
model	Fitted mhmm object
newdata	New stslist
Return value	most_probable_cluster()

Entry Parameters

Parameter	Required	Type	Description
model	✓	MHMM	Fitted model from build_mhmm() + fit_mhmm().
newdata	✗	SequenceData / None	Sequences to classify. Default: training data.
compress	✗	bool	Reuse likelihoods for repeated sequences when supported. Default False.

Returns

A NumPy array of cluster indices (0-based), one value per sequence.

For probabilistic membership, use posterior_probs_mhmm().

Example

Assume seq is a prepared SequenceData object with the observed sequences you want to model.

from sequenzo.seqhmm import build_mhmm, fit_mhmm, predict_mhmm

mhmm = build_mhmm(seq, n_clusters=3, n_states=4, random_state=42)
mhmm = fit_mhmm(mhmm)

labels = predict_mhmm(mhmm, compress=False)
print(labels)  # e.g. [0, 2, 1, 0, ...]

R Counterpart

• Closest R function: seqHMM most_probable_cluster()
• Mapping note: Returns integer cluster indices; map to cluster_names on the model if you set them at build time.

Notes

• Requires a fitted model.
• Hard assignments ignore uncertainty; prefer posterior_probs_mhmm() when sequences sit near cluster boundaries.

See Also

• Markov Chain Models Introduction maps the full HMM-family workflow.
• Model Comparison helps choose between fitted models.
• Sequenzo and seqHMM Mapping gives the R correspondence.

Authors

Code: Yuqi Liang

Documentation: Yuqi Liang

References

Helske, S., & Helske, J. (2019). Mixture hidden Markov models for sequence data: The seqHMM package in R. Journal of Statistical Software, 88(3), 1–32.