This package, LSAMon-R, implements a software monad for the specification of Latent Semantic Analysis (LSA) workflows.
LSAMon-R's computations are based on the document-term matrix creation functions in the package
SparseMatrixRecommender
and the packages for matrix factorization
irlba and
NonNegativeMatrixFactorization.
Here is an example pipeline:
lsaObj <-
LSAMonUnit( textHamlet ) %>%
LSAMonMakeDocumentTermMatrix( stemWordsQ = TRUE, stopWords = stopwords::stopwords() ) %>%
LSAMonApplyTermWeightFunctions( "IDF", "None", "Cosine" ) %>%
LSAMonExtractTopics( numberOfTopics = numberOfTopics, minNumberOfDocumentsPerTerm = minNumberOfDocumentsPerTerm, maxSteps = maxSteps, method = "NNMF" ) %>%
LSAMonEchoTopicsTable( numberOfTerms = 20 )To install use the command:
install.packages("irlba", "magrittr", "purrr")
devtools::install_github("antononcube/R-packages", subdir = "SparseMatrixRecommender")
devtools::install_github("antononcube/R-packages", subdir = "NonNegativeMatrixFactorization")
devtools::install_github("antononcube/R-packages", subdir = "LSAMon-R")