Subsequence Pattern Comparison Across Groups
Source:R/sequence_compare_htna.R
sequence_compare_htna.RdExtends Nestimate::sequence_compare() with a level argument that
lets the comparison run on type-level (meta-path) sequences as well
as state-level sequences.
Arguments
- x
A
netobject_group(from groupedbuild_network), anetobject(requiresgroup), or a wide-formatdata.frame(requiresgroup).- group
Character or vector. Column name or vector of group labels. Not needed for
netobject_group.- level
Character.
"state"(default) compares concrete state-level k-grams;"type"first replaces each state with its actor type so the comparison runs on meta-paths (e.g.Human->AI->Human). Only meaningful whenxis an htna network with an actor partition.- sub
Integer vector. Pattern lengths to analyze. Default:
3:5.- min_freq
Integer. Minimum frequency in each group for a pattern to be included. Default: 5.
- test
Character. Inference method: one of
"permutation"(default),"chisq", or"none". See Details.- iter
Integer. Permutation iterations. Only used when
test = "permutation". Default: 1000.- adjust
Character. P-value correction method (see
p.adjust). Default:"fdr".
Value
An object of class net_sequence_compare. See
Nestimate::sequence_compare() for full details and the
corresponding plot() method.
Details
Extracts all k-gram patterns (subsequences of length k) from the
sequences in each cohort, computes standardised residuals against
the independence model, and optionally runs a permutation or
chi-square test for differences in pattern rates between cohorts.
Operates on grouped htna networks (built via
build_htna(..., group = ...)), single htna networks paired with a
group argument, or wide-format sequence data with an explicit
group argument. The actor partition itself is not consumed when
level = "state" — sequence comparison is between cohorts of
sessions, not between actors. When level = "type", concrete codes
are folded into their actor type before pattern enumeration, so the
comparison runs on meta-paths.
See also
permutation_htna() for whole-network differences,
mosaic_plot_htna() for single-step transition residuals,
extract_meta_paths() for descriptive meta-path enumeration.
Examples
# \donttest{
data(human_ai)
grp <- build_htna(human_ai, actor_type = "actor_type", group = "phase")
#> Warning: A network with one long sequence is not recommended and can't be validated using bootstrap and other confirmatory testings.
#> Metadata aggregated per session: ties resolved by first occurrence in 'cluster' (24 sessions), 'actor_type' (9 sessions)
#> Warning: A network with one long sequence is not recommended and can't be validated using bootstrap and other confirmatory testings.
#> Metadata aggregated per session: ties resolved by first occurrence in 'session_date' (1 sessions), 'cluster' (18 sessions), 'actor_type' (15 sessions)
# State-level comparison (default)
sequence_compare_htna(grp, iter = 50)
#> Sequence Comparison [699 patterns | 2 groups: Early, Late]
#> Lengths: 3, 4, 5 | min_freq: 5 | permutation: 50 iter (fdr)
#>
#> pattern length freq_Early freq_Late
#> Specify->Request->Specify->Ask 4 10 58
#> Frustrate->Ask->Plan->Check 4 39 7
#> Specify->Request->Specify->Ask->Plan 5 8 35
#> Specify->Frustrate->Ask->Plan->Check 5 37 6
#> Specify->Request->Specify 3 82 116
#> Request->Specify->Frustrate->Ask->Plan 5 105 51
#> Ask->Plan->Frustrate 3 74 112
#> Request->Specify->Frustrate 3 160 88
#> Request->Specify->Ask 3 147 197
#> Execute->Ask->Request 3 6 22
#> prop_Early prop_Late resid_Early resid_Late effect_size p_value
#> 0.0010494281 0.0067963440 -6.296503 6.296503 8.420388 0.3114973
#> 0.0040927694 0.0008202484 4.356696 -4.356696 6.541789 0.3114973
#> 0.0008586455 0.0042062252 -4.500415 4.500415 6.420094 0.3114973
#> 0.0039712354 0.0007210672 4.369457 -4.369457 6.374350 0.3114973
#> 0.0084171628 0.0132601738 -3.194473 3.194473 5.130398 0.3114973
#> 0.0112697220 0.0061290710 3.640078 -3.640078 5.102890 0.3114973
#> 0.0075959762 0.0128029264 -3.542427 3.542427 5.033177 0.3114973
#> 0.0164237323 0.0100594422 3.756080 -3.756080 4.866214 0.3114973
#> 0.0150893040 0.0225194330 -3.733134 3.733134 4.556679 0.3114973
#> 0.0006158900 0.0025148605 -3.315483 3.315483 4.490829 0.3114973
#> ... and 689 more patterns
# Meta-path comparison
sequence_compare_htna(grp, level = "type", iter = 50)
#> Sequence Comparison [55 patterns | 2 groups: Early, Late]
#> Lengths: 3, 4, 5 | min_freq: 5 | permutation: 50 iter (fdr)
#>
#> pattern length freq_Early freq_Late prop_Early
#> Human->Human->Human->AI->Human 5 112 196 0.012021037
#> AI->Human->Human->Human->Human 5 41 71 0.004400558
#> Human->AI->AI->Human 4 976 796 0.102424179
#> Human->Human->AI->AI 4 846 646 0.088781614
#> Human->AI->AI 3 1203 986 0.123485937
#> Human->AI->Human 3 1813 1780 0.186101417
#> Human->AI->Human->AI->Human 5 647 695 0.069442954
#> AI->Human->AI->AI->AI 5 44 63 0.004722550
#> Human->Human->AI->AI->Human 5 676 525 0.072555544
#> AI->AI->Human->AI->Human 5 363 279 0.038961039
#> prop_Late resid_Early resid_Late effect_size p_value
#> 0.023554861 -5.837799 5.837799 7.627748 0.1540616
#> 0.008532628 -3.448798 3.448798 3.804741 0.1540616
#> 0.093273963 2.064052 -2.064052 3.517733 0.1540616
#> 0.075697211 3.189261 -3.189261 3.499731 0.1540616
#> 0.112711477 2.264188 -2.264188 3.290279 0.1540616
#> 0.203475080 -2.980988 2.980988 3.252581 0.1540616
#> 0.083523615 -3.520925 3.520925 2.749593 0.1540616
#> 0.007571205 -2.432177 2.432177 2.809446 0.1659125
#> 0.063093378 2.490338 -2.490338 2.556875 0.1659125
#> 0.033529624 1.922749 -1.922749 2.477371 0.1659125
#> ... and 45 more patterns
# }