embed_sentencespace: Build a Starspace model to be used for sentence embedding

View source: R/r-all-the-things.R

embed_sentencespaceR Documentation

Build a Starspace model to be used for sentence embedding

Description

Build a Starspace model to be used for sentence embedding

Usage

embed_sentencespace(
  x,
  model = "sentencespace.bin",
  early_stopping = 0.75,
  useBytes = FALSE,
  ...
)

Arguments

x

a data.frame with sentences containg the columns doc_id, sentence_id and token The doc_id is just an article or document identifier, the sentence_id column is a character field which contains words which are separated by a space and should not contain any tab characters

model

name of the model which will be saved, passed on to starspace

early_stopping

the percentage of the data that will be used as training data. If set to a value smaller than 1, 1-early_stopping percentage of the data which will be used as the validation set and early stopping will be executed. Defaults to 0.75.

useBytes

set to TRUE to avoid re-encoding when writing out train and/or test files. See writeLines for details

...

further arguments passed on to starspace except file, trainMode and fileFormat

Value

an object of class textspace as returned by starspace.

Examples


library(udpipe)
data(brussels_reviews_anno, package = "udpipe")
x <- subset(brussels_reviews_anno, language == "nl")
x$token <- x$lemma
x <- x[, c("doc_id", "sentence_id", "token")]
set.seed(123456789)
model <- embed_sentencespace(x, dim = 15, epoch = 15,
                             negSearchLimit = 1, maxNegSamples = 2)
plot(model)
sentences <- c("ook de keuken zijn zeer goed uitgerust .",
               "het appartement zijn met veel smaak inrichten en zeer proper .")
predict(model, sentences, type = "embedding")
starspace_embedding(model, sentences)

## Not run: 
library(udpipe)
data(dekamer, package = "ruimtehol")
x <- udpipe(dekamer$question, "dutch", tagger = "none", parser = "none", trace = 100)
x <- x[, c("doc_id", "sentence_id", "sentence", "token")]
set.seed(123456789)
model <- embed_sentencespace(x, dim = 15, epoch = 5, minCount = 5)
plot(model)
predict(model, "Wat zijn de cijfers qua doorstroming van 2016?",
        basedoc = unique(x$sentence))

embeddings <- starspace_embedding(model, unique(x$sentence), type = "document")
dim(embeddings)

sentence <- "Wat zijn de cijfers qua doorstroming van 2016?"
embedding_sentence <- starspace_embedding(model, sentence, type = "document")
mostsimilar <- embedding_similarity(embeddings, embedding_sentence)
head(sort(mostsimilar[, 1], decreasing = TRUE), 3)

## clean up for cran
file.remove(list.files(pattern = ".udpipe$"))

## End(Not run)

ruimtehol documentation built on May 29, 2024, 5:26 a.m.