crfsuite (version 0.4.2)

ner_download_modeldata: CRF Training data: download training data for doing Named Entity Recognition (NER)

Description

Download training data for doing Named Entity Recognition (NER)

Usage

ner_download_modeldata(
  type = c("conll2002-nl", "conll2002-es", "GermanNER", "wikiner-de-wp2",
    "wikiner-de-wp3", "wikiner-en-wp2", "wikiner-en-wp3", "wikiner-es-wp2",
    "wikiner-es-wp3", "wikiner-fr-wp2", "wikiner-fr-wp3", "wikiner-it-wp2",
    "wikiner-it-wp3", "wikiner-nl-wp2", "wikiner-nl-wp3", "wikiner-pl-wp3",
    "wikiner-pt-wp3", "wikiner-ru-wp2", "wikiner-ru-wp3"),
  docs = -Inf
)

Value

a data.frame with training data for a Named Entity Recognition task or an object of try-error in case of failure of downloading the data

Arguments

type

a character string with the type of data to download. See the function usage for all possible values. These data will be downloaded from either:

• NLTK-data forked repository: https://github.com/bnosac-dev/nltk_data/blob/gh-pages/packages/corpora/conll2002.zip

• FOX forked repository of GermanNER: https://github.com/bnosac-dev/FOX/tree/master/input/GermanNER

• FOX forked repository of WikiNER: https://github.com/bnosac-dev/FOX/tree/master/input/Wikiner

Please visit the information on these repositories first before you use these data in any commercial product.

docs

integer indicating how many documents to sample from the data (only used for data from the NLTK repository). This is only used to reduce CRAN R CMD check training time in the examples of this R package.

Examples

if (FALSE) {
x <- ner_download_modeldata("conll2002-nl")
x <- ner_download_modeldata("conll2002-es")
x <- ner_download_modeldata("GermanNER")
x <- ner_download_modeldata("wikiner-en-wp2")
x <- ner_download_modeldata("wikiner-nl-wp3")
x <- ner_download_modeldata("wikiner-fr-wp3")
}
## reduce number of docs
x <- ner_download_modeldata("conll2002-es", docs = 10)