A wrapper for RSpectra::svds that returns
a nicely-formatted latent semantic analysis topic model.
FitLsaModel(dtm, k, calc_coherence = TRUE, return_all = FALSE, ...)A document term matrix of class Matrix::dgCMatrix
Number of topics
Do you want to calculate probabilistic coherence of topics
after the model is trained? Defaults to TRUE.
Should all objects returned from RSpectra::svds be
returned here? Defaults to FALSE
Other arguments to pass to svds through
its opts parameter.
Returns a list with a minimum of three objects: phi,
theta, and sv. The rows of phi index topics and the
columns index tokens. The rows of theta index documents and the
columns index topics. sv is a vector of singular values.
Latent semantic analysis, LSA, uses single value decomposition to factor the document term matrix. In many LSA applications, TF-IDF weights are applied to the DTM before model fitting. However, this is not strictly necessary.
# Load a pre-formatted dtm
data(nih_sample_dtm)
# Convert raw word counts to TF-IDF frequency weights
idf <- log(nrow(nih_sample_dtm) / Matrix::colSums(nih_sample_dtm > 0))
dtm_tfidf <- Matrix::t(nih_sample_dtm) * idf
dtm_tfidf <- Matrix::t(dtm_tfidf)
# Fit an LSA model
model <- FitLsaModel(dtm = dtm_tfidf, k = 5)
str(model)
#> List of 6
#> $ sv : num [1:5] 181 156 150 144 143
#> $ theta : num [1:100, 1:5] 0.0213 0.0103 0.0093 0.0198 0.0144 ...
#> ..- attr(*, "dimnames")=List of 2
#> .. ..$ : chr [1:100] "8693991" "8693362" "8607498" "8697008" ...
#> .. ..$ : chr [1:5] "t_1" "t_2" "t_3" "t_4" ...
#> $ phi : num [1:5, 1:5210] 0.000263 -0.000897 0.000831 -0.000494 -0.000135 ...
#> ..- attr(*, "dimnames")=List of 2
#> .. ..$ : chr [1:5] "t_1" "t_2" "t_3" "t_4" ...
#> .. ..$ : chr [1:5210] "folding" "tosuprttedprtmnt" "importation" "hd" ...
#> $ gamma : num [1:5, 1:5210] 1.45e-06 -5.75e-06 5.54e-06 -3.43e-06 -9.45e-07 ...
#> ..- attr(*, "dimnames")=List of 2
#> .. ..$ : chr [1:5] "t_1" "t_2" "t_3" "t_4" ...
#> .. ..$ : chr [1:5210] "folding" "tosuprttedprtmnt" "importation" "hd" ...
#> $ coherence: Named num [1:5] 0.937 0.937 0.231 0.268 0.986
#> ..- attr(*, "names")= chr [1:5] "t_1" "t_2" "t_3" "t_4" ...
#> $ data :Formal class 'dgCMatrix' [package "Matrix"] with 6 slots
#> .. ..@ i : int [1:14073] 94 51 73 13 29 16 89 98 80 63 ...
#> .. ..@ p : int [1:5211] 0 1 2 3 4 5 6 7 8 9 ...
#> .. ..@ Dim : int [1:2] 100 5210
#> .. ..@ Dimnames:List of 2
#> .. .. ..$ : chr [1:100] "8693991" "8693362" "8607498" "8697008" ...
#> .. .. ..$ : chr [1:5210] "folding" "tosuprttedprtmnt" "importation" "hd" ...
#> .. ..@ x : num [1:14073] 4.61 4.61 4.61 4.61 4.61 ...
#> .. ..@ factors : list()
#> - attr(*, "class")= chr "lsa_topic_model"