RuleBasedNLP_JavaSentence: Interface to rules-based NLP (Java pipeline) to obtain...

In wlktan/LireNLPSystem: LireNLPSystem

Description

Interface to rules-based NLP (Java pipeline) to obtain sentence level predictions.

Usage

RuleBasedNLP_JavaSentence(segmented.reports, imageid = "imageid",
  bodyText = "body", impressionText = "impression",
  findings_longstring = paste(c("spondylolisthesis", "annular_fissure",
  "disc_bulge", "disc_degeneration", "disc_desiccation",
  "disc_height_loss", "disc_protrusion", "facet_degeneration"), collapse =
  ";"))

Arguments

segmented.reports	The input data frame
imageid	String of column that uniquely identifies each image
bodyText	String of column that indexes body of report text
impressionText	String of column that indexes impression section of report text
findings_longstring	String of findings separated by ";"

Value

A "long" data frame with the columns Finding imageid: ID that indexes reports Sentence: Exact string of the sentence Section of sentence: Whether the sentence was in body or impression regex: 1 if there is a keyword in the sentence, otherwise 0 negex: 1 if the keyword is negated, otherwise 0 **NOTE**: regex cannot be 1 if regex = 0 keyword: The keyword that triggered regex = 1

Examples

RuleBasedNLP_JavaSentence(df, findings_longstring = "facet_degeneration;disc_degeneration")

wlktan/LireNLPSystem documentation built on May 27, 2019, 12:13 p.m.