Package 'svs'

Title: Tools for Semantic Vector Spaces
Description: Various tools for semantic vector spaces, such as correspondence analysis (simple, multiple and discriminant), latent semantic analysis, probabilistic latent semantic analysis, non-negative matrix factorization, latent class analysis, EM clustering, logratio analysis and log-multiplicative (association) analysis. Furthermore, there are specialized distance measures, plotting functions and some helper functions.
Authors: Koen Plevoets [aut, cre]
Maintainer: Koen Plevoets <[email protected]>
License: GPL-3
Version: 3.1.1
Built: 2025-02-20 03:43:49 UTC
Source: https://github.com/cran/svs

Help Index

Tools for Semantic Vector Spaces

Description

This package offers various tools for semantic vector spaces. There are techniques for correspondence analysis (simple, multiple and discriminant), latent semantic analysis, probabilistic latent semantic analysis, non-negative matrix factorization, latent class analysis, EM clustering, logratio analysis and log-multiplicative (association) analysis. Furthermore, the package has specialized distance measures and plotting functions as well as some helper functions.

Contents

This package contains the following raw data files (in the folder extdata):

SndT_Fra.txt

Seventeen Dutch source words and their French translations.

SndT_Eng.txt

Seventeen Dutch source words and their English translations.

InvT_Fra.txt

Seventeen Dutch target words and their French source words.

InvT_Eng.txt

Seventeen Dutch target words and their English source words.

Ctxt_Dut.txt

Context words for seventeen Dutch words.

Ctxt_Fra.txt

Context words for seventeen Dutch words translated from French.

Ctxt_Eng.txt

Context words for seventeen Dutch words translated from English.

The (fast procedures for the) techniques in this package are:

fast_sca

Simple correspondence analysis.

fast_mca

Multiple correspondence analysis.

fast_dca

Discriminant correspondence analysis.

fast_lsa

Latent semantic analysis.

fast_psa

Probabilistic latent semantic analysis.

fast_nmf

Non-negative matrix factorization.

fast_lca

Latent class analysis.

fast_E_M

EM clustering.

fast_lra

Logratio analysis.

fast_lma

Log-multiplicative (association) analysis.

The complete overview of local and global weighting functions in this package can be found on weighting_functions.

The specialized distance measures are:

dist_chisquare

Chi-square distance.

dist_cosine

Cosine distance.

dist_wrt

Distance with respect to a certain point.

dist_wrt_centers

Distance with respect to cluster centers.

The specialized plotting functions are:

cd_plot

Cumulative distribution plot.

pc_plot

Parallel coordinate plot.

There are two helper functions for correspondence analysis:

freq_ca

Compute level frequencies (for a factor).

centers_ca

Compute coordinates for cluster centers.

There is one helper function for pvclust:

complete_pvpick

Complete the output of pvpick.

There is one helper function for igraph:

layout4bipartite

Create a layout matrix for a bipartite graph.

The remaining helper functions in this package are:

rep4dat

Repeat the rows of a data frame according to a frequency column.

vec2ddc

Transform a vector into a double-coded matrix.

dat2ddc

Transform a data frame into a double-coded matrix.

vec2ind

Transform a vector into an indicator matrix.

tab2dat

Transform a table into a data frame.

tab2ind

Transform a table into an indicator matrix.

dat2ind

Transform a data frame into an indicator matrix.

outerec

Recursive application of the outer product.

pmi

Pointwise mutual information.

MI

Mutual information.

log_or_0

Logarithmic transform.

Further reference

  • Many packages contain correspondence analysis: ca, FactoMineR, MASS and others.

  • For latent semantic analysis there is also the package lsa.

  • The package NMF provides more flexibility for non-negative matrix factorization.

  • For topic models there are the packages lda and topicmodels.

  • Latent class analysis can also be run in the package poLCA.

  • For log-ratio analysis there is also the package easyCODA.

  • The package gnm offers much flexibility for association analysis, i.e. fitting log-multiplicative or Goodman's RC models.

Link

As from 2023, this package is part of Module 10: Multivariate data analysis with R of the Summer School Methods in Language Sciences.

Author

Koen Plevoets, [email protected]

Acknowledgements

This package has benefited greatly from the helpful comments of Lore Vandevoorde, Pauline De Baets and Gert De Sutter. Thanks to Kurt Hornik, Uwe Ligges and Brian Ripley for their valuable recommendations when proofing this package.

Plotting a Cumulative Distribution

Description

A function for plotting a cumulative distribution.

Usage

cd_plot(
  x,
  inc = 0.01,
  col = "darkgrey",
  cex = 1,
  font = 1,
  family = "",
  srt = -45,
  pch = 20,
  pcol = "black",
  pbg = "white",
  pcex = cex,
  lcol = col,
  lwd = 1,
  lty = 1,
  xlim = NULL,
  ylim = NULL,
  xlab = NULL,
  ylab = NULL,
  main = NULL,
  sub = NULL
)

Arguments

x

A numeric vector.
inc

The (numeric) increment for constructing the sequence from 0 to ceiling(max(x)), plotted on the horizontal axis.
col

The color of the line and the text labels: see colors.
cex

The character expansion factor: a numberic value to specify the size of the text labels.
font

The font of the text labels: 1 for plain, 2 for bold, 3 for italic, and 4 for bold italic.
family

The font family of the text labels: "serif", "sans", "mono", or one of the Hershey fonts.
srt

The rotation angle (in degrees) of the text labels.
pch

The plotting character for displaying points: see points.
pcol

The color of the plotting character: see colors.
pbg

The background color of the plotting character: see colors.
pcex

The character expansion factor of the plotting character: a numeric value to specify the size of the plotting character.
lcol

The color of the line: see colors.
lwd

The line width of the line: a numeric value to specify the width of the line.
lty

The line type of the line: 0 or "blank", 1 or "solid", 2 or "dashed", 3 or "dotted", 4 or "dotdash", 5 or "longdash", 6 or "twodash".
xlim

A vector of two numeric values specifying the lower and upper limit between which to plot the horizontal axis.
ylim

A vector of two numeric values specifying the lower and upper limit between which to plot the vertical axis.
xlab

A character string for labelling of the horizontal axis.
ylab

A character string for labelling of the vertical axis.
main

A character string for the main title of the plot.
sub

A character string for the subtitle of the plot.

Value

A cumulative distribution plot.

Examples

SndT_Fra <- read.table(system.file("extdata", "SndT_Fra.txt", package = "svs"),
   header = TRUE, sep = "\t", quote = "\"", encoding = "UTF-8",
   stringsAsFactors = FALSE)
sca_SndT_Fra <- fast_sca(SndT_Fra)
dis_SndT_Fra <- dist_wrt(sca_SndT_Fra$pos1)
cd_plot(dis_SndT_Fra)

Compute Coordinates for Cluster Centers

Description

A helper function for computing the coordinates of cluster centers (typically used in correspondence analysis).

Usage

centers_ca(x, clusters, freq)

Arguments

x

A numeric matrix.
clusters

A clustering of the row levels of x: either a list or the output of kmeans.
freq

An optional vector of frequency counts for the row levels of x.

Value

A matrix containing the coordinates of the cluster centers.

See Also

freq_ca.

Examples

SndT_Fra <- read.table(system.file("extdata", "SndT_Fra.txt", package = "svs"),
   header = TRUE, sep = "\t", quote = "\"", encoding = "UTF-8",
   stringsAsFactors = FALSE)
sca_SndT_Fra <- fast_sca(SndT_Fra)
kcl_SndT_Fra <- kmeans(sca_SndT_Fra$pos1, centers = 7)
centers_ca(sca_SndT_Fra$pos1, clusters = kcl_SndT_Fra, freq = freq_ca(SndT_Fra[, 1]))

Complete the Output of pvpick

Description

A helper function to add the missing singleton clusters in the output of pvpick (from the package pvclust).

Usage

complete_pvpick(clusters, labels)

Arguments

clusters

A clustering by a call to pvpick.
labels

A character vector containing the exhaustive set of levels.

Value

A list with the singleton clusters inserted at the end (so that the set of clusters is exhaustive).

Context Words for seventeen Dutch Words

Description

The frequency table of seventeen Dutch synonyms of beginnen ("to begin") and their context words (from the Dutch Parallel Corpus).

Format

A table with 17 rows and 1404 columns.

Examples

Ctxt_Dut <- read.table(system.file("extdata", "Ctxt_Dut.txt", package = "svs"),
   header = TRUE, sep = "\t", quote = "\"", encoding = "UTF-8")
sca_Ctxt_Dut <- fast_sca(data.matrix(Ctxt_Dut))
sca_Ctxt_Dut
lsa_Ctxt_Dut <- fast_lsa(data.matrix(Ctxt_Dut))
lsa_Ctxt_Dut

Context Words for seventeen Dutch Words Translated from French

Description

The frequency table of seventeen Dutch synonyms of beginnen ("to begin") and their context words in texts translated from English (from the Dutch Parallel Corpus).

Format

A table with 17 rows and 609 columns.

Examples

Ctxt_Eng <- read.table(system.file("extdata", "Ctxt_Eng.txt", package = "svs"),
   header = TRUE, sep = "\t", quote = "\"", encoding = "UTF-8")
sca_Ctxt_Eng <- fast_sca(data.matrix(Ctxt_Eng))
sca_Ctxt_Eng
lsa_Ctxt_Eng <- fast_lsa(data.matrix(Ctxt_Eng))
lsa_Ctxt_Eng

Context Words for seventeen Dutch Words Translated from French

Description

The frequency table of seventeen Dutch synonyms of beginnen ("to begin") and their context words in texts translated from French (from the Dutch Parallel Corpus).

Format

A table with 17 rows and 612 columns.

Examples

Ctxt_Fra <- read.table(system.file("extdata", "Ctxt_Fra.txt", package = "svs"),
   header = TRUE, sep = "\t", quote = "\"", encoding = "UTF-8")
sca_Ctxt_Fra <- fast_sca(data.matrix(Ctxt_Fra))
sca_Ctxt_Fra
lsa_Ctxt_Fra <- fast_lsa(data.matrix(Ctxt_Fra))
lsa_Ctxt_Fra

Transform a Data Frame into a Double-Coded Matrix

Description

A helper function for transforming all columns of a data frame into a double-coded matrix.

Usage

dat2ddc(dat, limits = NULL)

dat2dc(dat, limits = NULL)

Arguments

dat

A data frame containing only numeric vectors or ordered factors as columns.
limits

A vector, matrix or list specifying the lower and upper limits for the columns in dat. A single vector of length two applies the same limits to all columns. The matrix or list must either have the same number of columns/components as the number of the number of columns in dat or they must have names. The names can be a subset of the column names of dat. In any case, the limits for each column must be a vector of length two, i.e. the matrix must have two rows or each list component should be a vector of length two. If left unspecified (the default), then the minimum and maximum of each column in dat will be taken as (separate) limits.

Value

A matrix with doubly-coded columns.

See Also

vec2ddc.

Transform a Data Frame into an Indicator Matrix

Description

A helper function for transforming a data frame into an indicator matrix.

Usage

dat2ind(dat, add_names = TRUE)

Arguments

dat

A data frame.
add_names

Logical specifying whether to add rownames to the resulting indicator matrix.

Value

An indicator matrix.

See Also

vec2ind.

Compute Chi-square Distances

Description

A function for computing chi-square distances.

Usage

dist_chisquare(x, diag = FALSE, upper = FALSE)

dist_chisq(x, diag = FALSE, upper = FALSE)

Arguments

x

A numeric matrix (containing coordinates).
diag

Logical specifying whether the diagonal of the resulting distance matrix should be printed.
upper

Logical specifying whether the upper triangle of the resulting distance matrix should be printed.

Value

A distance matrix.

Examples

SndT_Fra <- read.table(system.file("extdata", "SndT_Fra.txt", package = "svs"),
   header = TRUE, sep = "\t", quote = "\"", encoding = "UTF-8",
   stringsAsFactors = FALSE)
tab_SndT_Fra <- table(SndT_Fra)
dist_chisquare(tab_SndT_Fra)

Compute Cosine Distances

Description

A function for computing cosine distances.

Usage

dist_cosine(x, diag = FALSE, upper = FALSE)

dist_cos(x, diag = FALSE, upper = FALSE)

Arguments

x

A numeric matrix (containing coordinates).
diag

Logical specifying whether the diagonal of the resulting distance matrix should be printed.
upper

Logical specifying whether the upper triangle of the resulting distance matrix should be printed.

Details

The cosine distance equals 1 - the cosine similarity.

Value

A distance matrix.

Examples

SndT_Fra <- read.table(system.file("extdata", "SndT_Fra.txt", package = "svs"),
   header = TRUE, sep = "\t", quote = "\"", encoding = "UTF-8",
   stringsAsFactors = FALSE)
lsa_SndT_Fra <- fast_lsa(SndT_Fra)
dist_cosine(lsa_SndT_Fra$pos1[, 1:7])

Compute Distances with respect to a certain Point

Description

A function for computing (euclidean) distances with respect to a certain specified point.

Usage

dist_wrt(x, wrt = NULL)

Arguments

x

A numeric matrix (containing coordinates).
wrt

A specification of the point with respect to which to compute all distances: can be either a vector or the character label of one of the row levels in x. If NULL or NA, then the origin (i.e. the point c(0 , 0, 0,... )) is taken as the value.

Value

A matrix (containing distances between the rows of x and wrt).

Examples

SndT_Fra <- read.table(system.file("extdata", "SndT_Fra.txt", package = "svs"),
   header = TRUE, sep = "\t", quote = "\"", encoding = "UTF-8",
   stringsAsFactors = FALSE)
sca_SndT_Fra <- fast_sca(SndT_Fra)
dist_wrt(sca_SndT_Fra$pos1, wrt = "beginnen")

Compute Distances with respect to Cluster Centers

Description

A function for computing (euclidean) distances with respect to specified cluster centers.

Usage

dist_wrt_centers(x, clusters, freq = NULL, members_only = TRUE)

Arguments

x

A numeric matrix (containing coordinates).
clusters

A clustering of the row levels of x: either a list or the output of kmeans.
freq

An optional vector of frequency counts for the row levels of x.
members_only

Logical specifying whether the distances from the cluster centers should only be computed for the cluster members.

Value

A list with a matrix of distances for every cluster.

See Also

centers_ca, freq_ca.

Examples

SndT_Fra <- read.table(system.file("extdata", "SndT_Fra.txt", package = "svs"),
   header = TRUE, sep = "\t", quote = "\"", encoding = "UTF-8",
   stringsAsFactors = FALSE)
sca_SndT_Fra <- fast_sca(SndT_Fra)
kcl_SndT_Fra <- kmeans(sca_SndT_Fra$pos1, centers = 7)
dist_wrt_centers(sca_SndT_Fra$pos1, clusters = kcl_SndT_Fra, freq = freq_ca(SndT_Fra[, 1]))

Discriminant Correspondence Analysis

Description

A fast procedure for computing discriminant correspondence analysis.

Usage

fast_dca(dat, clusters1 = NULL, clusters2 = NULL, members = FALSE)

Arguments

dat

Input data: can be a table or a data frame (but the data frame must have only two columns).
clusters1

A clustering of the first set of levels: either a list or the output of kmeans.
clusters2

A clustering of the second set of levels: either a list or the output of kmeans.
members

Logical indicating whether the (supplementary) coordinates for the individual levels should also be computed.

Value

A list with components:

val

The eigenvalues or principal inertias, indicating how much each latent axis explains.
cen1

The coordinates of the cluster centers for the first set of levels.
cen2

The coordinates of the cluster centers for the second set of levels.
mem1

If members = TRUE: The coordinates of the first set of individual levels.
mem2

If members = TRUE: The coordinates of the second set of individual levels.

References

Abdi, H. (2007) Discriminant correspondence analysis. In: N. Salkind (ed.) Encyclopedia of measurement and statistics. Thousand Oaks: SAGE.

Examples

SndT_Fra <- read.table(system.file("extdata", "SndT_Fra.txt", package = "svs"),
   header = TRUE, sep = "\t", quote = "\"", encoding = "UTF-8",
   stringsAsFactors = FALSE)
sca_SndT_Fra <- fast_sca(SndT_Fra)
kcl_SndT_Fra <- kmeans(sca_SndT_Fra$pos1, centers = 7)
dca_SndT_Fra <- fast_dca(SndT_Fra, clusters1 = kcl_SndT_Fra)
dca_SndT_Fra

EM clustering

Description

A fast procedure for Expectation-Maximization clustering.

Usage

fast_E_M(dat, k, tol = 1e-08)

fast_EM(dat, k, tol = 1e-08)

Arguments

dat

Input data: can be a table or a data frame (but the data frame must have only two columns).
k

Numeric specification of the number of latent classes to compute.
tol

Numeric specification of the convergence criterion.

Details

This function assumes that the rows of a frequency table come from a multinomial distribution. The prior probabilities of the latent classes are initialized with a Dirichlet distribution (by means of rdirichlet from the package gtools) with alpha = the total frequency counts of every level.

Value

A list with components:

prob0

The probabilities of the latent classes.
prob1

The probabilities for the first set of levels (viz. the row levels of a frequency table). The rows of prob1 sum to 1.
prob2

The probabilities for the second set of levels (viz. the column levels of a frequency table). The rows of prob2 sum to 1.

References

Dempster, A. P., N. M. Laird and D. B. Rubin (1977) Maximum likelihood from incomplete data via the EM algorithm. Journal of the royal statistical society, series B 39 (1), 1–38.

Examples

SndT_Fra <- read.table(system.file("extdata", "SndT_Fra.txt", package = "svs"),
   header = TRUE, sep = "\t", quote = "\"", encoding = "UTF-8",
   stringsAsFactors = FALSE)
E_M_SndT_Fra <- fast_E_M(SndT_Fra, k = 7)
E_M_SndT_Fra

Latent Class Analysis

Description

A fast procedure for computing latent class analysis.

Usage

fast_lca(dat, k, tol = 1e-08, posterior = FALSE, transform = 1, sep = "_")

Arguments

dat

Input data: can be a table or a data frame.
k

Numeric specification of the number of latent classes to compute.
tol

Numeric specification of the convergence criterion.
posterior

Logical indicating whether the posterior probabilities of the individual observations should also be returned.
transform

Numeric specification of the power transformation to be applied on the posterior distribution; see fast_psa.
sep

Character specifying the separator string for joining the levels (if posterior = TRUE).

Details

The prior probabilities of the latent classes are initialized with a Dirichlet distribution (by means of rdirichlet from the package gtools) with alpha = the total frequency counts of every level.

Value

A list with components:

prob0

The probabilities of the latent classes.
prob1-prob...

The probabilities for each set of levels. The columns of each prob... sum to 1.
posterior

If posterior = TRUE: An indicator matrix with the posterior probabilities of each observation.

References

Agresti, A. (2013) Categorical data analysis. Hoboken: John Wiley and Sons, 535–542.

Examples

SndT_Fra <- read.table(system.file("extdata", "SndT_Fra.txt", package = "svs"),
   header = TRUE, sep = "\t", quote = "\"", encoding = "UTF-8",
   stringsAsFactors = FALSE)
lca_SndT_Fra <- fast_lca(SndT_Fra, k = 7)
lca_SndT_Fra

Log-Multiplicative Association Analysis

Description

A fast procedure for computing log-multiplicative analysis, i.e. Goodman's _RC(M)_ association model.

Usage

fast_lma(
  dat,
  k,
  weights = "marginal",
  tol = 1e-08,
  base = exp(1),
  init = "marginal"
)

fast_rca(
  dat,
  k,
  weights = "marginal",
  tol = 1e-08,
  base = exp(1),
  init = "marginal"
)

Arguments

dat

Input data: can be a table or a data frame.
k

Numeric specification of the number of latent axes to compute (i.e. k = M).
weights

Character specification of the weights applied to standardize the coordinates: can be one of "marginal", "uniform", "unit" or "none".
tol

Numeric specification of the convergence criterion.
base

Numeric specification of the base with respect to which logarithms are computed.
init

Character specification of the initialization scheme for the marginal parameters: can be either "kateri" or "marginal". This argument may change in future versions of the svs package.

Details

For now (i.e. version 3.0.0 of the svs package), the data frame must have only two columns.

Value

A list with components:

mar

A list with marginal parameters in components mar1 and mar2; not so important for the analysis.
val

The association parameters, indicating how much association each latent axis explains.
pos1

The coordinates of the first set of levels (viz. the row levels of a frequency table).
pos2

The coordinates of the second set of levels (viz. the column levels of a frequency table).

References

Goodman, L. A. (1979) Simple models for the analysis of association in cross-classifications having ordered categories. Journal of the American statistical association 74 (367), 537–552.

Kateri, M. (2014) Contingency table analysis. Methods and implementation using R. New York: Springer-Birkhauser.

Wong, R. S.-K. (2010) Association models. Thousand Oaks: SAGE.

Examples

SndT_Fra <- read.table(system.file("extdata", "SndT_Fra.txt", package = "svs"),
   header = TRUE, sep = "\t", quote = "\"", encoding = "UTF-8",
   stringsAsFactors = FALSE)
lma_SndT_Fra <- fast_lma(SndT_Fra, k = 7)
lma_SndT_Fra

Logratio Analysis

Description

A fast procedure for computing logratio analysis.

Usage

fast_lra(dat, base = exp(1))

Arguments

dat

Input data: can be a table or a data frame (but the data frame must have only two columns).
base

Numeric specification of the base with respect to which logarithms are computed.

Value

A list with components:

val

The eigenvalues (i.e. squared singular values), indicating how much each latent axis explains.
pos1

The coordinates of the first set of levels (viz. the row levels of a frequency table).
pos2

The coordinates of the second set of levels (viz. the column levels of a frequency table).

References

Greenacre, M. (2019) Compositional data analysis in practice. Boca Raton: Chapman and Hall/CRC.

Van den Boogaart, K. G. and R. Tolosana-Delgado (2013) Analyzing compositional data with R. Berlin: Springer.

Examples

SndT_Fra <- read.table(system.file("extdata", "SndT_Fra.txt", package = "svs"),
   header = TRUE, sep = "\t", quote = "\"", encoding = "UTF-8",
   stringsAsFactors = FALSE)
lra_SndT_Fra <- fast_lra(SndT_Fra)
lra_SndT_Fra

Latent Semantic Analysis

Description

A fast procedure for computing latent semantic analysis.

Usage

fast_lsa(dat, local_weights = "log", global_weights = "idf")

fast_lsi(dat, local_weights = "log", global_weights = "idf")

Arguments

dat

Input data: can be a table or a data frame (but the data frame must have only two columns).
local_weights

Character specification of the local weighting function (without a prefix): see Weighting functions.
global_weights

Character specification of the global weighting function (without a prefix): see Weighting functions.

Value

A list with components:

val

The singular values, indicating how much each latent axis explains.
pos1

The coordinates of the first set of levels (viz. the row levels of a frequency table).
pos2

The coordinates of the second set of levels (viz. the column levels of a frequency table).

References

Deerwester, S., S. T. Dumais, G. W. Furnas, Th. K. Landauer and R. Harshman (1990) Indexing by latent semantic analysis. Journal of the American society for information science 41 (6), 391–407.

Landauer, Th. K. and S. T. Dumais (1997) A solution to Plato's problem: the latent semantic analysis theory of the acquisition, induction, and representation of knowledge. Psychological review 104, 211–240.

Examples

SndT_Fra <- read.table(system.file("extdata", "SndT_Fra.txt", package = "svs"),
   header = TRUE, sep = "\t", quote = "\"", encoding = "UTF-8",
   stringsAsFactors = FALSE)
lsa_SndT_Fra <- fast_lsa(SndT_Fra)
lsa_SndT_Fra

Multiple Correspondence Analysis

Description

A fast procedure for computing multiple correspondence analysis.

Usage

fast_mca(dat, nfac = FALSE)

Arguments

dat

Input data: has to be a data frame (with any number of columns).
nfac

Logical indicating whether the number of factors (i.e. the number of columns in dat) is a divisor for the eigenvalues (principal inertias) and the coordinates.

Value

A list with components:

val

The eigenvalues or principal inertias, indicating how much each latent axis explains.
pos

The coordinates of all levels.

References

Greenacre, M. (2017) Correspondence analysis in practice, Third edition. Boca Raton: Chapman and Hall/CRC.

Examples

SndT_Fra <- read.table(system.file("extdata", "SndT_Fra.txt", package = "svs"),
   header = TRUE, sep = "\t", quote = "\"", encoding = "UTF-8",
   stringsAsFactors = FALSE)
mca_SndT_Fra <- fast_mca(SndT_Fra)
mca_SndT_Fra

Non-negative Matrix Factorization

Description

A fast procedure for non-negative matrix factorization.

Usage

fast_nmf(dat, k, type = "KL", tol = 1e-08)

fast_nmf_KL(dat, k, tol = 1e-08)

fast_nmf_Fr(dat, k, tol = 1e-08)

fast_nmf_Al(dat, k, tol = 1e-08)

Arguments

dat

Input data: can be a table or a data frame (but the data frame must have only two columns).
k

Numeric specification of the number of latent axes to compute.
type

Character specification of the type of optimization: can in the current implementation be either "KL" for the Kullback-Leibler divergence, "Frobenius" or "euclidean" (or abbreviations thereof) for the euclidean distance, or "ALS" for alternating least squares.
tol

Numeric specification of the convergence criterion.

Value

A list with components:

pos1

The coordinates of the first set of levels (viz. the row levels of a frequency table).
pos2

The coordinates of the second set of levels (viz. the column levels of a frequency table).

References

Lee, D. D. and H. S. Seung (1999) Learning the parts of objects by non-negative matrix factorization. Nature 401, 788–791.

Lee, D. D. and H. S. Seung (2001) Algorithms for non-negative matrix factorization. Advances in neural information processing systems 13, 556–562.

Examples

SndT_Fra <- read.table(system.file("extdata", "SndT_Fra.txt", package = "svs"),
   header = TRUE, sep = "\t", quote = "\"", encoding = "UTF-8",
   stringsAsFactors = FALSE)
nmf_SndT_Fra <- fast_nmf(SndT_Fra, k = 7)
nmf_SndT_Fra

Probabilistic Latent Semantic Analysis

Description

A fast procedure for computing probabilistic latent semantic analysis.

Usage

fast_psa(dat, k, symmetric = FALSE, transform = 1, tol = 1e-08)

fast_psi(dat, k, symmetric = FALSE, transform = 1, tol = 1e-08)

fast_plsa(dat, k, symmetric = FALSE, transform = 1, tol = 1e-08)

fast_plsi(dat, k, symmetric = FALSE, transform = 1, tol = 1e-08)

Arguments

dat

Input data: can be a table or a data frame (but the data frame must have only two columns).
k

Numeric specification of the number of latent classes to compute.
symmetric

Logical indicating whether to compute the symmetric or the asymmetric solution.
transform

Numeric specification of the "tempering" transformation as explained in Hofmann (1999: 51-52).
tol

Numeric specification of the convergence criterion.

Details

From version 1.1.0 of the svs package on, probabilistic latent semantic analysis is a special case of latent class analysis.

Value

A list with components:

prob0

The probabilities of the latent classes.
prob1

The probabilities for the first set of levels (viz. the row levels of a frequency table). The rows of prob1 sum to 1 if symmetric = FALSE, the columns sum to 1 if symmetric = TRUE.
prob2

The probabilities for the second set of levels (viz. the column levels of a frequency table). The columns of prob2 sum to 1.

References

Hofmann, Th. (1999). Probabilistic latent semantic indexing. SIGIR'99: Proceedings of the 22nd annual international SIGIR conference on research and development in information retrieval, 50–57.

Examples

SndT_Fra <- read.table(system.file("extdata", "SndT_Fra.txt", package = "svs"),
   header = TRUE, sep = "\t", quote = "\"", encoding = "UTF-8",
   stringsAsFactors = FALSE)
psa_SndT_Fra <- fast_psa(SndT_Fra, k = 7)
psa_SndT_Fra

Simple Correspondence Analysis

Description

A fast procedure for computing simple correspondence analysis.

Usage

fast_sca(dat, transform = 1)

Arguments

dat

Input data: can be a table or a data frame (but the data frame must have only two columns).
transform

Numeric specification of the power transformation to be applied on the data.

Value

A list with components:

val

The eigenvalues or principal inertias, indicating how much each latent axis explains.
pos1

The coordinates of the first set of levels (viz. the row levels of a frequency table).
pos2

The coordinates of the second set of levels (viz. the column levels of a frequency table).

References

Greenacre, M. (2017) Correspondence analysis in practice, Third edition. Boca Raton: Chapman and Hall/CRC.

Examples

SndT_Fra <- read.table(system.file("extdata", "SndT_Fra.txt", package = "svs"),
   header = TRUE, sep = "\t", quote = "\"", encoding = "UTF-8",
   stringsAsFactors = FALSE)
sca_SndT_Fra <- fast_sca(SndT_Fra)
sca_SndT_Fra

Compute Level Frequencies (for a Factor or Vector)

Description

A helper function for computing the frequency of each factor level (typically used in correspondence analysis).

Usage

freq_ca(dat, nfac = FALSE)

Arguments

dat

A factor, (character) vector or a data frame.
nfac

Logical indicating whether the number of factors (i.e. the number of columns in dat) is a divisor for the level frequencies.

Value

A vector containing the frequency counts of every level.

Examples

SndT_Fra <- read.table(system.file("extdata", "SndT_Fra.txt", package = "svs"),
   header = TRUE, sep = "\t", quote = "\"", encoding = "UTF-8",
   stringsAsFactors = FALSE)
freq_ca(SndT_Fra)

Seventeen Dutch Target Words and their English Source Words

Description

The occurrences of seventeen Dutch synonyms of beginnen ("to begin") and their English source words (from the Dutch Parallel Corpus).

Format

A data frame with 782 rows and 2 variables.

source_Eng

The English source word.

target_Dut

The Dutch target word.

Examples

InvT_Eng <- read.table(system.file("extdata", "InvT_Eng.txt", package = "svs"),
   header = TRUE, sep = "\t", quote = "\"", encoding = "UTF-8",
   stringsAsFactors = FALSE)
sca_InvT_Eng <- fast_sca(InvT_Eng)
sca_InvT_Eng
lsa_InvT_Eng <- fast_lsa(InvT_Eng)
lsa_InvT_Eng

Seventeen Dutch Target Words and their French Source Words

Description

The occurrences of seventeen Dutch synonyms of beginnen ("to begin") and their French source words (from the Dutch Parallel Corpus).

Format

A data frame with 856 rows and 2 variables.

source_Fra

The French source word.

target_Dut

The Dutch target word.

Examples

InvT_Fra <- read.table(system.file("extdata", "InvT_Fra.txt", package = "svs"),
   header = TRUE, sep = "\t", quote = "\"", encoding = "UTF-8",
   stringsAsFactors = FALSE)
sca_InvT_Fra <- fast_sca(InvT_Fra)
sca_InvT_Fra
lsa_InvT_Fra <- fast_lsa(InvT_Fra)
lsa_InvT_Fra

Create a Layout Matrix for a Bipartite Graph

Description

A helper function for creating a layout matrix for a bipartite graph, based on the dimensions of its biadjacency matrix.

Usage

layout4bipartite(nrow, ncol, rowsLeft = TRUE)

layout4diagram(nrow, ncol, rowsLeft = TRUE)

Arguments

nrow

Numeric specifying the number of rows of the biadjacency matrix.
ncol

Numeric specifying the number of columns of the biadjacency matrix.
rowsLeft

Logical specifying whether the rows of the biadjacency matrix should appear to the left of the columns.

Details

This is essentially an alternative for layout_as_bipartite() from the igraph package. Bipartite graphs are typically represented by a biadjacency matrix, the dimensions of which can be used for constructing the layout. In addition, the coordinates for the vertices are somewhat different from those in igraph.

Value

A matrix.

Logarithmic transform

Description

A function for computing the logarithm of every entry in a table with 0 for zero entries.

Usage

log_or_0(x, base = exp(1))

Arguments

x

A table or a (sparse) matrix.
base

Numeric specification of the base with respect to which logarithms are computed.

Value

An array with the logarithm of every entry and 0 for all zero entries.

Examples

SndT_Fra <- read.table(system.file("extdata", "SndT_Fra.txt", package = "svs"),
   header = TRUE, sep = "\t", quote = "\"", encoding = "UTF-8",
   stringsAsFactors = FALSE)
tab_SndT_Fra <- table(SndT_Fra)
log_or_0(tab_SndT_Fra)

Mutual Information

Description

A function for computing the mutual information.

Usage

MI(x, base = 2)

mi(x, base = 2)

Arguments

x

A table or a (sparse) matrix.
base

Numeric specification of the base with respect to which logarithms are computed.

Value

A numeric value containing the mutual information.

See Also

pmi.

Examples

SndT_Fra <- read.table(system.file("extdata", "SndT_Fra.txt", package = "svs"),
   header = TRUE, sep = "\t", quote = "\"", encoding = "UTF-8",
   stringsAsFactors = FALSE)
tab_SndT_Fra <- table(SndT_Fra)
MI(tab_SndT_Fra)

Recursive Application of the Outer Product

Description

A helper function for computing the outer product of two or more arrays.

Usage

outerec(...)

Arguments

...

The specification of two or more arrays (separated by comma's or contained in a list).

Value

An array with the outer product of all the arrays specified in ....

Plotting Parallel Coordinates

Description

A function for plotting parallel coordinates.

Usage

pc_plot(
  x,
  col = "darkgrey",
  cex = 1,
  font = 1,
  family = "",
  pch = 20,
  pcol = col,
  pcex = cex,
  lcol = col,
  lwd = 1,
  lty = 1,
  acol = "black",
  alwd = 1,
  alty = 1,
  las = 1,
  add_scale = FALSE,
  main = NULL,
  sub = NULL
)

Arguments

x

A numeric matrix.
col

The color of the text labels, points and connecting lines: see colors.
cex

The character expansion factor: A numeric value to specify the size of the text labels and the points.
font

The font of the text labels: 1 for plain, 2 for bold, 3 for italic, and 4 for bold italic.
family

The font family of the text labels: "serif", "sans", "mono", or one of the Hershey fonts.
pch

The plotting character for displaying points: see points.
pcol

The color of the plotting character: see colors.
pcex

The character expansion factor of the plotting character: a numeric value to specify the size of the plotting character.
lcol

The color of the connecting lines: see colors.
lwd

The line width of the connecting lines: a numeric value to specify the width of the connecting lines.
lty

The line type of the connecting lines: 0 or "blank", 1 or "solid", 2 or "dashed", 3 or "dotted", 4 or "dotdash", 5 or "longdash", 6 or "twodash".
acol

The color of the parallel axes: see colors.
alwd

The line width of the parallel axes: a numeric value to specify the width of the parallel axes.
alty

The line type of the parallel axes: 0 or "blank", 1 or "solid", 2 or "dashed", 3 or "dotted", 4 or "dotdash", 5 or "longdash", 6 or "twodash".
las

The reading direction of the labels on the axes ("label axis style"): either a numeric value between 0 and 3 (see las in par), or a character value matching either "horizontal" or "vertical".
add_scale

Logical specifying whether to add a scale for the parallel axes (which are normalized).
main

A character string for the main title of the plot.
sub

A character string for the subtitle of the plot.

Value

A parallel coordinate plot.

Examples

SndT_Fra <- read.table(system.file("extdata", "SndT_Fra.txt", package = "svs"),
   header = TRUE, sep = "\t", quote = "\"", encoding = "UTF-8",
   stringsAsFactors = FALSE)
sca_SndT_Fra <- fast_sca(SndT_Fra)
pc_plot(sca_SndT_Fra$pos1, las = "vertical")

Pointwise Mutual Information

Description

A function for computing the pointwise mutual information of every entry in a table.

Usage

pmi(x, normalize = FALSE, base = 2)

PMI(x, normalize = FALSE, base = 2)

Arguments

x

A table or a (sparse) matrix.
normalize

Logical indicating whether to normalize the pointwise mutual information.
base

Numeric specification of the base with respect to which logarithms are computed.

Value

An array with the pointwise mutual information of every entry.

See Also

MI.

Examples

SndT_Fra <- read.table(system.file("extdata", "SndT_Fra.txt", package = "svs"),
   header = TRUE, sep = "\t", quote = "\"", encoding = "UTF-8",
   stringsAsFactors = FALSE)
tab_SndT_Fra <- table(SndT_Fra)
pmi(tab_SndT_Fra)

Repeat the Rows of a Data Frame according to a Frequency Column

Description

A helper function for repeating the rows a data frame according to a frequency column.

Usage

rep4dat(dat, freqName = "Freq")

Arguments

dat

A data frame.
freqName

Character specifying the name of the frequency column.

Value

A data frame.

Examples

SndT_Fra <- read.table(system.file("extdata", "SndT_Fra.txt", package = "svs"),
   header = TRUE, sep = "\t", quote = "\"", encoding = "UTF-8",
   stringsAsFactors = FALSE)
tab_SndT_Fra <- table(SndT_Fra)
dat_SndT_Fra <- as.data.frame(tab_SndT_Fra)
rep4dat(dat_SndT_Fra)

Seventeen Dutch Source Words and their English Translations

Description

The occurrences of seventeen Dutch synonyms of beginnen ("to begin") and their English translations (from the Dutch Parallel Corpus).

Format

A data frame with 1117 rows and 2 variables.

source_Dut

The Dutch source word.

target_Eng

The English target word.

Examples

SndT_Eng <- read.table(system.file("extdata", "SndT_Eng.txt", package = "svs"),
   header = TRUE, sep = "\t", quote = "\"", encoding = "UTF-8",
   stringsAsFactors = FALSE)
sca_SndT_Eng <- fast_sca(SndT_Eng)
sca_SndT_Eng
lsa_SndT_Eng <- fast_lsa(SndT_Eng)
lsa_SndT_Eng

Seventeen Dutch Source Words and their French Translations

Description

The occurrences of seventeen Dutch synonyms of beginnen ("to begin") and their French translations (from the Dutch Parallel Corpus).

Format

A data frame with 1487 rows and 2 variables.

source_Dut

The Dutch source word.

target_Fra

The French target word.

Examples

SndT_Fra <- read.table(system.file("extdata", "SndT_Fra.txt", package = "svs"),
   header = TRUE, sep = "\t", quote = "\"", encoding = "UTF-8",
   stringsAsFactors = FALSE)
sca_SndT_Fra <- fast_sca(SndT_Fra)
sca_SndT_Fra
lsa_SndT_Fra <- fast_lsa(SndT_Fra)
lsa_SndT_Fra

Transform a Table into a Data Frame

Description

A helper function for transforming a table into a data frame.

Usage

tab2dat(tab)

Arguments

tab

A table or (sparse) matrix.

Value

A data frame.

Transform a Table into an Indicator Matrix

Description

A helper function for transforming a table into an indicator matrix.

Usage

tab2ind(tab, add_names = TRUE)

Arguments

tab

A table or (sparse) matrix.
add_names

Logical specifying whether to add rownames to the resulting indicator matrix.

Value

An indicator matrix.

Transform a Vector into a Double-Coded Matrix

Description

A helper function for transforming a vector into a double-coded matrix.

Usage

vec2ddc(x, limits = c(NA, NA))

vec2dc(x, limits = c(NA, NA))

Arguments

x

A vector, either numeric or an ordered factor.
limits

Numeric vector of length two specifying the limits (or "poles") of the underlying ordinal variable. If left unspecified (the default), then the minimum and maximum in x will be taken.

Details

Double coding, also called data doubling, is a way of indicating how the values of an ordinal variable relate to the lower and upper limits of the variable. Columns - and + express how close each (observed) value is to these respective limits.

Value

A matrix with column names - and +.

References

Greenacre, M. (2014) Data doubling and fuzzy coding. In: J. Blasius and M. Greenacre (eds.) Visualization and verbalization of data. Chapman and Hall/CRC, 239–253. Greenacre, M. (2017) Correspondence analysis in practice, Third edition. Boca Raton: Chapman and Hall/CRC, 201–208.

See Also

dat2ddc.

Transform a Vector into an Indicator Matrix

Description

A helper function for transforming a vector into an indicator matrix.

Usage

vec2ind(x, add_names = TRUE)

Arguments

x

A vector (which will internally be converted to a factor).
add_names

Logical specifying whether to add dimnames to the resulting indicator matrix.

Details

As of version 2.0.x of the svs package, this is essentially a wrapper for t(fac2sparse()) from the Matrix package.

Value

An indicator matrix.

See Also

dat2ind.

Weighting Functions

Description

Local and global weighting functions.

Usage

lw_tf(x)

lw_raw(x)

lw_log(x)

lw_bin(x)

gw_idf(x)

gw_idf_alt(x)

gw_gfidf(x)

gw_nor(x)

gw_ent(x)

gw_bin(x)

gw_raw(x)

Arguments

x

A numeric matrix.

Details

There are many local and global weighting functions. In this package, local weighting functions are prefixed with lw_ and global weighting functions with gw_, so users can define their own weighting functions.

Local weighting functions (i.e. weighting every cell in the matrix):

lw_tf

Term frequency: f(x) = x.

lw_raw

Raw frequency, which is the same as the term frequency: f(x) = x.

lw_log

Logarithm: f(x) = log(x + 1).

lw_bin

Binary: f(x) = 1 if x > 0 and 0 otherwise.

Global weighting functions, weighting the columns of the matrix (hence, these weighting functions work according to expectation for a document-term matrix, i.e. with the documents as the rows and the terms as the columns):

gw_idf

Inverse document frequency: f(x) = log( nrow(x) / n + 1) where n = the number of rows in which the column >0.

gw_idf_alt

Alternative definition of the inverse document frequency: f(x) = log( nrow(x) / n) + 1 where n = the number of rows in which the column >0.

gw_gfidf

Global frequency multiplied by inverse document frequency: f(x) = colSums(x) / n where n = the number of rows in which the column >0.

gw_nor

Normal(ized) frequency: f(x) = x / colSums(x^2).

gw_ent

Entropy: f(x) = 1 + the relative Shannon entropy.

gw_bin

Binary: f(x) = 1.

gw_raw

Raw, which is the same as binary: f(x) = 1.

Value

A numeric matrix.

See Also

fast_lsa.

Examples

SndT_Fra <- read.table(system.file("extdata", "SndT_Fra.txt", package = "svs"),
   header = TRUE, sep = "\t", quote = "\"", encoding = "UTF-8",
   stringsAsFactors = FALSE)
tab_SndT_Fra <- table(SndT_Fra)
lw_log(tab_SndT_Fra)
gw_idf(tab_SndT_Fra)