get_lrt_test()

get_lrt_test() compares two predefined sequence groups and returns only the likelihood ratio test (LRT), not BIC.

Function Usage

get_lrt_test(
    seqdata,
    seqdata2=None,
    group=None,
    set_var=None,
    s=100,
    seed=36963,
    squared="LRTonly",
    weighted=True,
    opt=None,
    BFopt=None,
    method="OM",
    **kwargs
)

TraMineR Parameter Mapping

• seqdata, seqdata2, group, set_var -> TraMineRextras seqLRT data/group inputs
• s, seed, squared, weighted, opt -> seqLRT control arguments

Entry Parameters

Parameter	Required	Type	Description
seqdata	✓	SequenceData / DataFrame / list	Main input sequences.
seqdata2	✗	SequenceData / DataFrame / list / None	Optional second sequence input.
group	✗	array-like / None	Two-group labels used when splitting one dataset.
set_var	✗	array-like / None	Optional stratification variable.
s	✗	int	Resampling size (0 means full comparison).
seed	✗	int	Random seed.
squared	✗	bool / str	If string, only "LRTonly" is valid.
weighted	✗	bool / str	Use weights (True/False) or "by.group".
opt	✗	int / None	Internal sampling option.
BFopt	✗	int / None	Reserved Bayes-factor option passed through the engine.
method	✗	str	Distance method (for example "OM").
**kwargs	✗	any	Extra distance-method arguments.

What It Returns

A NumPy array with two columns:

• LRT
• p-value

If set_var is provided, each row corresponds to one stratum.

Example

from sequenzo.group_comparison import get_lrt_test

lrt = get_lrt_test(
    seqdata=seqdata,
    group=df["group"],
    method="OM",
    indel=1,
    sm="CONSTANT"
)

print(lrt)

R Counterpart

• Closest R function: TraMineRextras seqLRT
• Mapping note: LRT-only wrapper of the same two-group comparison engine.

Notes

• Internally this function calls get_group_differences(..., stat="LRT").
• The same input constraints apply: exactly two groups among valid cases.
• group and seqdata2 cannot both be missing.

Authors

Code: Yuqi Liang

Documentation: Yuqi Liang

References

Liao, T. F., & Fasang, A. E. (2021). Comparing groups of life-course sequences using the Bayesian information criterion and the likelihood-ratio test. Sociological Methodology, 51(1), 44-85.