Skip to contents

Pairwise Similarity or Distance Matrix

sim_matrix.Rd

Calculate a matrix of similarity or scores between the rows of the input.

Usage

sim_matrix(x, ...)

# S3 method for class 'data.frame'
sim_matrix(
  x,
  cols,
  method = c("cosine", "cosine_squished", "euclidean", "minkowski", "dot_prod",
    "anchored"),
  ...,
  tidy_output = FALSE
)

Arguments

x

an embeddings object, matrix, or dataframe with one embedding per row

...

additional parameters to be passed to method function

cols

tidyselect - columns that contain numeric embedding values

method

either the name of a method to compute similarity or distance, or a function that takes two vectors, x and y, and outputs a scalar, similar to those listed in Similarity and Distance Metrics

tidy_output

logical. If FALSE (the default), output a stats::dist object. If TRUE, output a tibble with columns doc_id_1, doc_id_2, and the similarity or distance metric.

Details

Available Methods

When method is the name of one of the following supported methods, computations are done with matrix operations and are therefore blazing fast.

  • cosine: cosine similarity

  • cosine_squished: cosine similarity, rescaled to range from 0 to 1

  • euclidean: Euclidean distance

  • minkowski: Minkowski distance; requires parameter p. When p = 1 (the default), this is the Manhattan distance. When p = 2, it is the Euclidean distance. When p = Inf, it is the Chebyshev distance.

  • dot_prod: Dot product

When method is a custom function, operations are performed for each row and may be slow for large inputs.

Value

If tidy_output = FALSE (the default), a stats::dist object. If tidy_output = TRUE, a tibble with columns doc_id_1, doc_id_2, and the similarity or distance metric.

Examples

emb <- predict(glove_twitter_25d, c("table", "chair", "cat"))

sim_matrix(emb)
#>           table     chair
#> chair 0.8680218          
#> cat   0.7297673 0.7455769
sim_matrix(emb, method = "euclidean")
#>          table    chair
#> chair 2.421414         
#> cat   3.300149 3.286518
sim_matrix(emb, method = function(x, y) sum(abs(x - y)))
#>          table    chair
#> chair 10.33815         
#> cat   13.02696 13.05754

valence_df <- tibble::as_tibble(emb, rownames = "token")
valence_df |> sim_matrix(dim_1:dim_25, tidy_output = TRUE)
#> # A tibble: 3 × 3
#>   doc_id_1 doc_id_2 cosine   
#>      <int>    <int> <dist>   
#> 1        1        2 0.8680218
#> 2        1        3 0.7297673
#> 3        2        3 0.7455769