Extract a non directional graph based on co-occurrence in the token. It extracts only if two entities are mentioned in the same token (sentence or paragraph)

extract_graph(
  text,
  using = "sentences",
  connect = connectors("misc"),
  sw = gen_stopwords("en"),
  loop = FALSE
)

Arguments

text: an input text
using: sentence or paragraph to tokenize
connect: lowercase connectors, like the "von" in "John von Neumann".
sw: stopwords vector.
loop: if TRUE, it will not remove loops, a node pointing to itself.

Examples

text <- "John Does lives in New York in United States of America. He  is a passionate jazz musician, often playing in local clubs."
extract_graph(text)
#> Tokenizing by sentences
#> # A tibble: 6 × 2
#>   n1                       n2                      
#>   <chr>                    <chr>                   
#> 1 John Does                New York                
#> 2 John Does                United States of America
#> 3 John Does                He                      
#> 4 New York                 United States of America
#> 5 New York                 He                      
#> 6 United States of America He