Linguistic Matching and Accommodation

Measure similarity between texts. Offers a variety of processing tools and similarity metrics to facilitate flexible representation of texts and matching. Implements forms of Language Style Matching (Ireland & Pennebaker, 2010) and Latent Semantic Analysis (Landauer & Dumais, 1997) .

lingmatch

An all-in-one R package for the assessment of linguistic matching and/or accommodation.

features

• Input raw text, a document-term matrix (DTM), or LIWC output.
• Apply various weighting functions to a DTM.
• Measure similarity and/or accommodation with various metrics.
• Calculate standard forms of Language Style Matching (LSM) and Latent Semantic Similarity (LSS).

resources

• Documentation and guides: miserman.github.io/lingmatch
  • Quick Start
  • Comparison Specification
  • Introduction to Text Analysis
  • Word Vectors
  • Text Classification
  • Dictionary Creation
• Dictionary repository: osf.io/y6g5b
• Latent semantic space repository: osf.io/489he
• Dictionary builder: miserman.github.io/dictionary_builder

installation

Download R from r-project.org, then install the package from an R console:

Release (version 1.0.7)

install.packages("lingmatch")

Development (version 1.0.8)

# install.packages("remotes")
remotes::install_github("miserman/lingmatch")

And load the package:

library(lingmatch)

examples

Can make a quick comparison between two bits of text; by default this will give the cosine similarity between raw word-count vectors:

lingmatch("First text to look at.", "Text to compare that text with.")

Or, given a vector of texts:

text = c(
  "Why, hello there! How are you this evening?",
  "I am well, thank you for your inquiry!",
  "You are a most good at social interactions person!",
  "Why, thank you! You're not all bad yourself!"
)

Process the texts in one step:

# with a dictionary
inquirer_cats = lma_process(text, dict = "inquirer", dir = "~/Dictionaries")

# with a latent semantic space
glove_vectors = lma_process(text, space = "glove", dir = "~/Latent Semantic Spaces")

Or process the texts step by step, then measure similarity between each:

dtm = lma_dtm(text)
dtm_weighted = lma_weight(dtm)
dtm_categorized = lma_termcat(dtm_weighted, lma_dict(1:9))
similarity = lma_simets(dtm_categorized, metric = "canberra")

Or do that within a single function call:

similarity = lingmatch(
  text, weight = "frequency", dict = lma_dict(1:9), metric = "canberra"
)$sim

Or, if you want a standard form (as in this example), specify a default:

similarity = lingmatch(text, type = "lsm")$sim