plot_coincidences_annotators_pairwise_old: Cross-plots of tags from a pair of annotators, faceted by...

In petereckley/webannor: Import text annotation files exported by WebAnno

Description

Pairwise coccurrences of topics on the same document, from any annotator, including multiple topic tags applied by same annotator to same document. A given tag will be counted (m-1) times, where m is the number of tags applied by any annotator on a given document, i.e. once for each pair of tags that the tag appears in. Note that to the extent that there are multiple tags from the same annotator (on a given document) with the tags applied by different annotators, this plot does *not* provide a completely clean measure of agreement between annotators (it conflates cooccurence within annotator).

Usage

plot_coincidences_annotators_pairwise_old(annotations,
  annotators = get_annotators(annotations), tagset_planned = NULL,
  margins = TRUE, cutoff = 0, sort = c("value", "frequency"))

plot_coincidences_annotators_pairwise(annotations,
  annotators = get_annotators(annotations), tagset_planned = NULL,
  margins = TRUE, cutoff = 0, sort = c("value", "frequency"))

Arguments

annotations	data.frame containing annotations data
annotators	list of annotators between whom to count coincidences
tagset_planned	character vector of tag names offered to annotators
margins	Whether to add marginal plots
cutoff	Threshold number of coincidences per cell below which to leave cell blank
sort	Method to order the tagset on plot

Details

But I don't see a straightforward how to measure agreement where there are multiple tags per annotator.

Functions

• plot_coincidences_annotators_pairwise: Alternative version of the function while I try to work around a bug

petereckley/webannor documentation built on May 25, 2019, 12:48 a.m.