Learn R Programming
quallmer (version 0.4.0)

trail_compare: trail_compare: run a task across multiple settings and compute reliability (deprecated)

Description

[Deprecated]

Usage

trail_compare(
  data,
  text_col,
  task,
  settings,
  id_col = NULL,
  label_col = "label",
  cache_dir = NULL,
  overwrite = FALSE,
  annotate_fun = annotate,
  min_coders = 2L
)

Value

A trail_compare object with components:

records

Named list of trail_record objects (one per setting)

matrix

Wide coder-style annotation matrix (settings = columns)

icr

Named list of inter-rater reliability statistics

meta

Metadata on settings, identifiers, task, timestamp, etc.

Arguments

data

A data frame containing the text to be annotated.

text_col

Character scalar. Name of the text column containing text units to annotate.

task

A quallmer task object describing what to extract or label.

settings

A named list of trail_setting objects. The list names serve as identifiers for each setting (similar to coder IDs).

id_col

Optional character scalar identifying the unit column. If NULL, a consistent temporary ID (".trail_unit_id") is created and added to the input data so annotations from all settings can be aligned.

label_col

Character scalar. Name of the label column in each record's annotations data that should be used as the code for comparison (e.g. "label", "score", "category").

cache_dir

Optional character scalar specifying a directory to cache LLM outputs. Passed to trail_record(). If NULL, caching disabled. For examples and tests, use tempdir() to comply with CRAN policies.

overwrite

Logical. If TRUE, ignore all cached results and recompute annotations for every setting.

annotate_fun

Annotation backend function used by trail_record().

min_coders

Minimum number of non-missing coders per unit required for inclusion in the inter-rater reliability calculation.

Details

trail_compare() is deprecated. Use qlm_replicate() to re-run coding with different models or settings, then use qlm_compare() to assess inter-rater reliability.

All settings are applied to the same text units. Because the ID column is shared across settings, their annotation outputs can be directly compared via the matrix component, and summarized using inter-rater reliability statistics in icr.

See Also

  • trail_record() – run a task for a single setting

  • trail_matrix() – align records into coder-style wide format

  • trail_icr() – compute inter-rater reliability across settings