Title: | Two-Table ExPosition |
---|---|
Description: | An extension of ExPosition for two table analyses, specifically, discriminant analyses. |
Authors: | Derek Beaton, Jenny Rieck, Cherise R. Chin Fatt, Herve Abdi |
Maintainer: | Derek Beaton <[email protected]> |
License: | GPL-2 |
Version: | 2.6.10.1 |
Built: | 2024-10-27 06:26:57 UTC |
Source: | CRAN |
TExPosition is two-table ExPosition
and includes discriminant methods of the singular value decomposition (SVD). The core of TExPosition is ExPosition
and the svd
.
Package: | TExPosition |
Type: | Package |
Version: | 2.6.10 |
Date: | 2013-12-00 |
Depends: | R (>=2.15.0), prettyGraphs (>= 2.1.4), ExPosition (>= 2.0.0) |
License: | GPL-2 |
URL: | http://www.utdallas.edu/~derekbeaton/software/ExPosition |
Questions, comments, compliments, and complaints go to Derek Beaton [email protected].
The following people are authors or contributors to TExPosition code, data, or examples:
Derek Beaton, Jenny Rieck, Cherise Chin-Fatt, Francesca Filbey, and Hervé Abdi.
Abdi, H., and Williams, L.J. (2010). Principal component analysis. Wiley Interdisciplinary Reviews: Computational Statistics, 2, 433-459.
Abdi, H. and Williams, L.J. (2010). Correspondence analysis. In N.J. Salkind, D.M., Dougherty, & B. Frey (Eds.): Encyclopedia of Research Design. Thousand Oaks (CA): Sage. pp. 267-278.
Abdi, H. (2007). Singular Value Decomposition (SVD) and Generalized Singular Value Decomposition (GSVD). In N.J. Salkind (Ed.): Encyclopedia of Measurement and Statistics.Thousand Oaks (CA): Sage. pp. 907-912.
Abdi, H. & Williams, L.J. (2010). Barycentric discriminant analysis (BADIA). In N.J. Salkind, D.M., Dougherty, & B. Frey (Eds.): Encyclopedia of Research Design. Thousand Oaks (CA): Sage. pp. 64-75.
Abdi, H. (2007). Discriminant correspondence analysis. In N.J. Salkind (Ed.): Encyclopedia of Measurement and Statistics. Thousand Oaks (CA): Sage. pp. 270-275.
Krishnan, A., Williams, L. J., McIntosh, A. R., & Abdi, H. (2011). Partial Least Squares (PLS) methods for neuroimaging: A tutorial and review. NeuroImage, 56(2), 455 – 475.
McIntosh, A. R., & Lobaugh, N. J. (2004). Partial least squares analysis of neuroimaging data: applications and advances. Neuroimage, 23, S250–S263.
tepBADA
, tepPLS
, tepGPLS
, tepDICA
, tepPLSCA
#For more examples, see each individual function (as noted above).
#For more examples, see each individual function (as noted above).
Calculates constraints for plotting latent variables.
calculateLVConstraints(results,x_axis=1,y_axis=2,constraints=NULL)
calculateLVConstraints(results,x_axis=1,y_axis=2,constraints=NULL)
results |
results (with $lx and $ly) from TExPosition (i.e., $TExPosition.Data) |
x_axis |
which component should be on the x axis? |
y_axis |
which component should be on the y axis? |
constraints |
if available, axis constraints for the plots (determines end points of the plots). |
Returns a list with the following items:
$constraints |
axis constraints for the plots (determines end points of the plots). |
Derek Beaton
Fast Euclidean distance calculations.
fastEucCalc(x, c)
fastEucCalc(x, c)
x |
a set of points. |
c |
a set of centers. |
This function is especially useful for discriminant analyses. The distance from each point in x
to each point in c
is computed and returned as a nrow(x)
x nrow(c)
matrix.
a distance matrix |
Euclidean distances of each point to each center are returned. |
Hervé Abdi, Derek Beaton
All computations between individual factor scores (fii) and group factor scores (fi).
fii2fi(DESIGN, fii, fi)
fii2fi(DESIGN, fii, fi)
DESIGN |
a dummy-coded design matrix |
fii |
a set of factor scores for individuals (rows) |
fi |
a set of factor scores for rows |
A list of values containing:
distances |
Euclidean distances of all rows to each category center |
assignments |
an assignment matrix (similar to DESIGN) where each individual is assigned to the closest category center |
confusion |
a confusion matrix of how many items are assigned (and mis-assigned) to each category |
Hervé Abdi, Derek Beaton
Print assignment results.
## S3 method for class 'tepAssign' print(x,...)
## S3 method for class 'tepAssign' print(x,...)
x |
an list that contains items to make into the tepAssign class. |
... |
inherited/passed arguments for S3 print method(s). |
Derek Beaton, Cherise Chin-Fatt
Print tepBADA results.
## S3 method for class 'tepBADA' print(x,...)
## S3 method for class 'tepBADA' print(x,...)
x |
an list that contains items to make into the tepBADA class. |
... |
inherited/passed arguments for S3 print method(s). |
Derek Beaton, Cherise Chin-Fatt
Print tepDICA results.
## S3 method for class 'tepDICA' print(x,...)
## S3 method for class 'tepDICA' print(x,...)
x |
an list that contains items to make into the tepDICA class. |
... |
inherited/passed arguments for S3 print method(s). |
Derek Beaton, Cherise Chin-Fatt
Print tepGPLS results.
## S3 method for class 'tepGPLS' print(x,...)
## S3 method for class 'tepGPLS' print(x,...)
x |
an list that contains items to make into the tepGPLS class. |
... |
inherited/passed arguments for S3 print method(s). |
Derek Beaton, Cherise Chin-Fatt
Print tepGraphs results.
## S3 method for class 'tepGraphs' print(x,...)
## S3 method for class 'tepGraphs' print(x,...)
x |
an list that contains items to make into the tepGraphs class. |
... |
inherited/passed arguments for S3 print method(s). |
Derek Beaton, Cherise Chin-Fatt
Print tepPLS results.
## S3 method for class 'tepPLS' print(x,...)
## S3 method for class 'tepPLS' print(x,...)
x |
an list that contains items to make into the tepPLS class. |
... |
inherited/passed arguments for S3 print method(s). |
Derek Beaton, Cherise Chin-Fatt
Print tepPLSCA results.
## S3 method for class 'tepPLSCA' print(x,...)
## S3 method for class 'tepPLSCA' print(x,...)
x |
an list that contains items to make into the tepPLSCA class. |
... |
inherited/passed arguments for S3 print method(s). |
Derek Beaton, Cherise Chin-Fatt
Print TExPosition results.
## S3 method for class 'texpoOutput' print(x,...)
## S3 method for class 'texpoOutput' print(x,...)
x |
an list that contains items to make into the texpoOutput class. |
... |
inherited/passed arguments for S3 print method(s). |
Derek Beaton, Cherise Chin-Fatt
A function to compute R-squared for BADA and DICA
R2(group.masses, di, ind.masses = NULL, dii)
R2(group.masses, di, ind.masses = NULL, dii)
group.masses |
a masses matrix for the groups |
di |
a set of squared distances of the groups |
ind.masses |
a masses matrix for the individuals |
dii |
a set of squared distances for the individuals |
R2 |
An R-squared |
Jenny Rieck, Derek Beaton
Barycentric Discriminant Analysis (BADA) via TExPosition.
tepBADA(DATA, scale = TRUE, center = TRUE, DESIGN = NULL, make_design_nominal = TRUE, group.masses = NULL, weights = NULL, graphs = TRUE, k = 0)
tepBADA(DATA, scale = TRUE, center = TRUE, DESIGN = NULL, make_design_nominal = TRUE, group.masses = NULL, weights = NULL, graphs = TRUE, k = 0)
DATA |
original data to perform a BADA on. |
scale |
a boolean, vector, or string. See |
center |
a boolean, vector, or string. See |
DESIGN |
a design matrix to indicate if rows belong to groups. Required for BADA. |
make_design_nominal |
a boolean. If TRUE (default), DESIGN is a vector that indicates groups (and will be dummy-coded). If FALSE, DESIGN is a dummy-coded matrix. |
group.masses |
a diagonal matrix or column-vector of masses for the groups. |
weights |
a diagonal matrix or column-vector of weights for the column items. |
graphs |
a boolean. If TRUE (default), graphs and plots are provided (via |
k |
number of components to return. |
Note: BADA is a special case of PLS (tepPLS
,tepGPLS
) wherein DATA1 are data and DATA2 are a group-coded disjunctive matrix. This is also called mean-centered PLS (Krishnan et al., 2011).
See epGPCA
(and also corePCA
) for details on what is returned. In addition to the values returned:
fii |
factor scores computed for supplemental observations |
dii |
squared distances for supplemental observations |
rii |
cosines for supplemental observations |
assign |
|
lx |
latent variables from DATA1 computed for observations |
ly |
latent variables from DATA2 computed for observations |
Derek Beaton
Abdi, H., and Williams, L.J. (2010). Principal component analysis. Wiley Interdisciplinary Reviews: Computational Statistics, 2, 433-459.
Abdi, H. and Williams, L.J. (2010). Correspondence analysis. In N.J. Salkind, D.M., Dougherty, & B. Frey (Eds.): Encyclopedia of Research Design. Thousand Oaks (CA): Sage. pp. 267-278.
Abdi, H. (2007). Singular Value Decomposition (SVD) and Generalized Singular Value Decomposition (GSVD). In N.J. Salkind (Ed.): Encyclopedia of Measurement and Statistics.Thousand Oaks (CA): Sage. pp. 907-912.
Abdi, H. & Williams, L.J. (2010). Barycentric discriminant analysis (BADIA). In N.J. Salkind, D.M., Dougherty, & B. Frey (Eds.): Encyclopedia of Research Design. Thousand Oaks (CA): Sage. pp. 64-75.
Abdi, H., Williams, L.J., Beaton, D., Posamentier, M., Harris, T.S., Krishnan, A., & Devous, M.D. (in press, 2012). Analysis of regional cerebral blood flow data to discriminate among Alzheimer's disease, fronto-temporal dementia, and elderly controls: A multi-block barycentric discriminant analysis (MUBADA) methodology. Journal of Alzheimer Disease, , -.
Abdi, H., Williams, L.J., Connolly, A.C., Gobbini, M.I., Dunlop, J.P., & Haxby, J.V. (2012). Multiple Subject Barycentric Discriminant Analysis (MUSUBADA): How to assign scans to categories without using spatial normalization. Computational and Mathematical Methods in Medicine, 2012, 1-15. doi:10.1155/2012/634165.
Krishnan, A., Williams, L. J., McIntosh, A. R., & Abdi, H. (2011). Partial Least Squares (PLS) methods for neuroimaging: A tutorial and review. NeuroImage, 56(2), 455 – 475.
corePCA
, epPCA
, epGPCA
, epMDS
For MatLab code: http://utd.edu/~derekbeaton/attachments/Software/matlab/MuSuBADA_V3.zip
data(bada.wine) bada.res <- tepBADA(bada.wine$data,scale=FALSE,DESIGN=bada.wine$design,make_design_nominal=FALSE)
data(bada.wine) bada.res <- tepBADA(bada.wine$data,scale=FALSE,DESIGN=bada.wine$design,make_design_nominal=FALSE)
Discriminant Correspondence Analysis (DICA) via TExPosition.
tepDICA(DATA, make_data_nominal = FALSE, DESIGN = NULL, make_design_nominal = TRUE, group.masses = NULL, weights = NULL, symmetric = TRUE, graphs = TRUE, k = 0)
tepDICA(DATA, make_data_nominal = FALSE, DESIGN = NULL, make_design_nominal = TRUE, group.masses = NULL, weights = NULL, symmetric = TRUE, graphs = TRUE, k = 0)
DATA |
original data to perform a DICA on. Data can be contingency (like CA) or categorical (like MCA). |
make_data_nominal |
a boolean. If TRUE (default), DATA is recoded as a dummy-coded matrix. If FALSE, DATA is a dummy-coded matrix. |
DESIGN |
a design matrix to indicate if rows belong to groups. Required for DICA. |
make_design_nominal |
a boolean. If TRUE (default), DESIGN is a vector that indicates groups (and will be dummy-coded). If FALSE, DESIGN is a dummy-coded matrix. |
group.masses |
a diagonal matrix or column-vector of masses for the groups. |
weights |
a diagonal matrix or column-vector of weights for the column it |
symmetric |
a boolean. If TRUE (default) symmetric factor scores for rows. |
graphs |
a boolean. If TRUE (default), graphs and plots are provided (via |
k |
number of components to return. |
If you use Hellinger distance, it is best to set symmetric
to FALSE.
Note: DICA is a special case of PLS-CA (tepPLSCA
) wherein DATA1 are data and DATA2 are a group-coded disjunctive matrix.
See epCA
(and also coreCA
) for details on what is returned. In addition to the values returned:
fii |
factor scores computed for supplemental observations |
dii |
squared distances for supplemental observations |
rii |
cosines for supplemental observations |
assign |
|
lx |
latent variables from DATA1 computed for observations |
ly |
latent variables from DATA2 computed for observations |
Derek Beaton, Hervé Abdi
Abdi, H., and Williams, L.J. (2010). Principal component analysis. Wiley Interdisciplinary Reviews: Computational Statistics, 2, 433-459.
Abdi, H. and Williams, L.J. (2010). Correspondence analysis. In N.J. Salkind, D.M., Dougherty, & B. Frey (Eds.): Encyclopedia of Research Design. Thousand Oaks (CA): Sage. pp. 267-278.
Abdi, H. (2007). Singular Value Decomposition (SVD) and Generalized Singular Value Decomposition (GSVD). In N.J. Salkind (Ed.): Encyclopedia of Measurement and Statistics.Thousand Oaks (CA): Sage. pp. 907-912.
Abdi, H. (2007). Discriminant correspondence analysis. In N.J. Salkind (Ed.): Encyclopedia of Measurement and Statistics. Thousand Oaks (CA): Sage. pp. 270-275.
Pinkham, A.E., Sasson, N.J., Beaton, D., Abdi, H., Kohler, C.G., Penn, D.L. (in press, 2012). Qualitatively distinct factors contribute to elevated rates of paranoia in autism and schizophrenia. Journal of Abnormal Psychology, 121, -.
Williams, L.J., Abdi, H., French, R., & Orange, J.B. (2010). A tutorial on Multi-Block Discriminant Correspondence Analysis (MUDICA): A new method for analyzing discourse data from clinical populations. Journal of Speech Language and Hearing Research, 53, 1372-1393.
Williams, L.J., Dunlop, J.P., & Abdi, H. (2012). Effect of age on the variability in the production of text-based global inferences. PLoS One, 7(5): e36161. doi:10.1371/ journal.pone.0036161 (pp.1-9)
coreCA
, epCA
, epMCA
For MatLab code: http://utd.edu/~herve/HerveAbdi_MatlabPrograms4MUDICA.zip
For additional R code (with inference tests): http://utdallas.edu/~dfb090020/attachments/MuDiCA.zip
data(dica.wine) dica.res <- tepDICA(dica.wine$data,DESIGN=dica.wine$design,make_design_nominal=FALSE)
data(dica.wine) dica.res <- tepDICA(dica.wine$data,DESIGN=dica.wine$design,make_design_nominal=FALSE)
Generalized Partial Least Squares (GPLS) via TExPosition. GPLS is to PLS (tepPLS
) as PCA epPCA
is to GPCA epGPCA
.
The major difference between PLS and GPLS is that GPLS allows the use of weights for the columns of each data set (just like GPCA).
tepGPLS(DATA1, DATA2, center1 = TRUE, scale1 = "SS1", center2 = TRUE, scale2 = "SS1", DESIGN = NULL, make_design_nominal = TRUE, weights1 = NULL, weights2 = NULL, graphs = TRUE, k = 0)
tepGPLS(DATA1, DATA2, center1 = TRUE, scale1 = "SS1", center2 = TRUE, scale2 = "SS1", DESIGN = NULL, make_design_nominal = TRUE, weights1 = NULL, weights2 = NULL, graphs = TRUE, k = 0)
DATA1 |
Data matrix 1 (X) |
DATA2 |
Data matrix 2 (Y) |
center1 |
a boolean, vector, or string to center |
scale1 |
a boolean, vector, or string to scale |
center2 |
a boolean, vector, or string to center |
scale2 |
a boolean, vector, or string to scale |
DESIGN |
a design matrix to indicate if rows belong to groups. |
make_design_nominal |
a boolean. If TRUE (default), DESIGN is a vector that indicates groups (and will be dummy-coded). If FALSE, DESIGN is a dummy-coded matrix. |
weights1 |
a weight vector (or diag matrix) for the columns of DATA1. |
weights2 |
a weight vector (or diag matrix) for the columns of DATA2. |
graphs |
a boolean. If TRUE (default), graphs and plots are provided (via |
k |
number of components to return. |
This implementation of Partial Least Squares is a symmetric analysis. It was first described by Tucker (1958), again by Bookstein (1994), and has gained notoriety in Neuroimaging from McIntosh et al., (1996). This particular implementation allows the user to provide weights for the columns of both DATA1
and DATA2
.
See epGPCA
(and also corePCA
) for details on what is returned. In addition to the values returned:
lx |
latent variables from DATA1 computed for observations |
ly |
latent variables from DATA2 computed for observations |
data1.norm |
center and scale information for DATA1 |
data1.norm |
center and scale information for DATA2 |
Derek Beaton
Tucker, L. R. (1958). An inter-battery method of factor analysis. Psychometrika, 23(2), 111–136.
Bookstein, F., (1994). Partial least squares: a dose–response model for measurement in the behavioral and brain sciences. Psycoloquy 5 (23)
Abdi, H. (2007). Singular Value Decomposition (SVD) and Generalized Singular Value Decomposition (GSVD). In N.J. Salkind (Ed.): Encyclopedia of Measurement and Statistics.Thousand Oaks (CA): Sage. pp. 907-912.
Krishnan, A., Williams, L. J., McIntosh, A. R., & Abdi, H. (2011). Partial Least Squares (PLS) methods for neuroimaging: A tutorial and review. NeuroImage, 56(2), 455 – 475.
McIntosh, A. R., & Lobaugh, N. J. (2004). Partial least squares analysis of neuroimaging data: applications and advances. Neuroimage, 23, S250–S263.
corePCA
, epPCA
, epGPCA
, tepPLS
, tepPLSCA
, tepBADA
, tepDICA
data(beer.tasting.notes) data1<-beer.tasting.notes$data[,1:8] data2<-beer.tasting.notes$data[,9:16] gpls.res <- tepGPLS(data1,data2)
data(beer.tasting.notes) data1<-beer.tasting.notes$data[,1:8] data2<-beer.tasting.notes$data[,9:16] gpls.res <- tepGPLS(data1,data2)
TExPosition plotting function which is an interface to prettyGraphs
.
tepGraphs(res, x_axis = 1, y_axis = 2, tepPlotInfo = NULL, DESIGN = NULL, fi.col = NULL, fi.pch = NULL, fii.col = NULL, fii.pch = NULL, fj.col = NULL, fj.pch = NULL, col.offset = NULL, constraints = NULL, lv.constraints = NULL, xlab = NULL, ylab = NULL, main = NULL, lvPlots = TRUE, lvAgainst = TRUE, contributionPlots = TRUE, correlationPlotter = TRUE, showHulls = 1, biplots = FALSE, graphs = TRUE)
tepGraphs(res, x_axis = 1, y_axis = 2, tepPlotInfo = NULL, DESIGN = NULL, fi.col = NULL, fi.pch = NULL, fii.col = NULL, fii.pch = NULL, fj.col = NULL, fj.pch = NULL, col.offset = NULL, constraints = NULL, lv.constraints = NULL, xlab = NULL, ylab = NULL, main = NULL, lvPlots = TRUE, lvAgainst = TRUE, contributionPlots = TRUE, correlationPlotter = TRUE, showHulls = 1, biplots = FALSE, graphs = TRUE)
res |
results from TExPosition |
x_axis |
which component should be on the x axis? |
y_axis |
which component should be on the y axis? |
tepPlotInfo |
A list ( |
DESIGN |
A design matrix to apply colors (by pallete selection) to row items |
fi.col |
A matrix of colors for the group items. If NULL, colors will be selected. |
fi.pch |
A matrix of pch values for the group items. If NULL, pch values are all 21. |
fii.col |
A matrix of colors for the row items (observations). If NULL, colors will be selected. |
fii.pch |
A matrix of pch values for the row items (observations). If NULL, pch values are all 21. |
fj.col |
A matrix of colors for the column items. If NULL, colors will be selected. |
fj.pch |
A matrix of pch values for the column items. If NULL, pch values are all 21. |
col.offset |
A numeric offset value. Is passed to |
constraints |
Plot constraints as returned from |
lv.constraints |
Plot constraints for latent variables. If NULL, constraints are selected. |
xlab |
x axis label |
ylab |
y axis label |
main |
main label for the graph window |
lvPlots |
a boolean. If TRUE, latent variables (X, Y) are plotted. If FALSE, latent variables are not plotted. |
lvAgainst |
a boolean. If TRUE, latent variables (X, Y) are plotted against each other. If FALSE, latent variables are plotted like factor scores. |
contributionPlots |
a boolean. If TRUE (default), contribution bar plots will be created. |
correlationPlotter |
a boolean. If TRUE (default), a correlation circle plot will be created. Applies to PCA family of methods (CA is excluded for now). |
showHulls |
a value between 0 and 1 to make a peeled hull at that percentage. All values outside of 0-1 will not plot any hulls. |
biplots |
a boolean. If FALSE (default), separate plots are made for row items ($fii and $fi) and column items ($fj). If TRUE, row ($fii and $fi) and column ($fj) items will be on the same plot. |
graphs |
a boolean. If TRUE, graphs are created. If FALSE, only data associated to plotting (e.g., constraints, colors) are returned. |
tepGraphs is an interface between TExPosition
and prettyGraphs
.
The following items are bundled inside of $Plotting.Data:
$fii.col |
the colors that are associated to the individuals (row items; $fii). |
$fii.pch |
the pch values associated to the individuals (row items; $fii). |
$fi.col |
the colors that are associated to the groups ($fi). |
$fi.pch |
the pch values associated to the groups ($fi). |
$fj.col |
the colors that are associated to the column items ($fj). |
$fj.pch |
the pch values associated to the column items ($fj). |
$constraints |
axis constraints for the plots (determines end points of the plots). |
Derek Beaton
#this is for TExPosition's iris data data(ep.iris) bada.iris <- tepBADA(ep.iris$data,DESIGN=ep.iris$design,make_design_nominal=FALSE) #there are only 2 components, not 3. bada.iris.plotting.data.biplot <- tepGraphs(bada.iris,x_axis=1,y_axis=2,biplots=TRUE)
#this is for TExPosition's iris data data(ep.iris) bada.iris <- tepBADA(ep.iris$data,DESIGN=ep.iris$design,make_design_nominal=FALSE) #there are only 2 components, not 3. bada.iris.plotting.data.biplot <- tepGraphs(bada.iris,x_axis=1,y_axis=2,biplots=TRUE)
Partial Least Squares (PLS) via TExPosition.
tepPLS(DATA1, DATA2, center1 = TRUE, scale1 = "SS1", center2 = TRUE, scale2 = "SS1", DESIGN = NULL, make_design_nominal = TRUE, graphs = TRUE, k = 0)
tepPLS(DATA1, DATA2, center1 = TRUE, scale1 = "SS1", center2 = TRUE, scale2 = "SS1", DESIGN = NULL, make_design_nominal = TRUE, graphs = TRUE, k = 0)
DATA1 |
Data matrix 1 (X) |
DATA2 |
Data matrix 2 (Y) |
center1 |
a boolean, vector, or string to center |
scale1 |
a boolean, vector, or string to scale |
center2 |
a boolean, vector, or string to center |
scale2 |
a boolean, vector, or string to scale |
DESIGN |
a design matrix to indicate if rows belong to groups. |
make_design_nominal |
a boolean. If TRUE (default), DESIGN is a vector that indicates groups (and will be dummy-coded). If FALSE, DESIGN is a dummy-coded matrix. |
graphs |
a boolean. If TRUE (default), graphs and plots are provided (via |
k |
number of components to return. |
This implementation of Partial Least Squares is a symmetric analysis. It was first described by Tucker (1958), again by Bookstein (1994), and has gained notoriety in Neuroimaging from McIntosh et al., (1996).
See epGPCA
(and also corePCA
) for details on what is returned. In addition to the values returned:
lx |
latent variables from DATA1 computed for observations |
ly |
latent variables from DATA2 computed for observations |
data1.norm |
center and scale information for DATA1 |
data1.norm |
center and scale information for DATA2 |
Derek Beaton
Tucker, L. R. (1958). An inter-battery method of factor analysis. Psychometrika, 23(2), 111–136.
Bookstein, F., (1994). Partial least squares: a dose–response model for measurement in the behavioral and brain sciences. Psycoloquy 5 (23)
McIntosh, A. R., Bookstein, F. L., Haxby, J. V., & Grady, C. L. (1996). Spatial Pattern Analysis of Functional Brain Images Using Partial Least Squares. NeuroImage, 3(3), 143–157.
Krishnan, A., Williams, L. J., McIntosh, A. R., & Abdi, H. (2011). Partial Least Squares (PLS) methods for neuroimaging: A tutorial and review. NeuroImage, 56(2), 455 – 475.
McIntosh, A. R., & Lobaugh, N. J. (2004). Partial least squares analysis of neuroimaging data: applications and advances. Neuroimage, 23, S250–S263.
corePCA
, epPCA
, epGPCA
, tepBADA
, tepGPLS
, tepPLSCA
data(beer.tasting.notes) data1<-beer.tasting.notes$data[,1:8] data2<-beer.tasting.notes$data[,9:16] pls.res <- tepPLS(data1,data2)
data(beer.tasting.notes) data1<-beer.tasting.notes$data[,1:8] data2<-beer.tasting.notes$data[,9:16] pls.res <- tepPLS(data1,data2)
Partial Least Squares-Correspondence Analysis (PLSCA) via TExPosition.
tepPLSCA(DATA1, DATA2, make_data1_nominal = FALSE, make_data2_nominal = FALSE, DESIGN = NULL, make_design_nominal = TRUE, weights1=NULL, weights2 = NULL, symmetric = TRUE, graphs = TRUE, k = 0)
tepPLSCA(DATA1, DATA2, make_data1_nominal = FALSE, make_data2_nominal = FALSE, DESIGN = NULL, make_design_nominal = TRUE, weights1=NULL, weights2 = NULL, symmetric = TRUE, graphs = TRUE, k = 0)
DATA1 |
Data matrix 1 (X), must be categorical (like MCA) or in disjunctive code see |
DATA2 |
Data matrix 2 (Y), must be categorical (like MCA) or in disjunctive code see |
make_data1_nominal |
a boolean. If TRUE (default), DATA1 is recoded as a dummy-coded matrix. If FALSE, DATA1 is a dummy-coded matrix. |
make_data2_nominal |
a boolean. If TRUE (default), DATA2 is recoded as a dummy-coded matrix. If FALSE, DATA2 is a dummy-coded matrix. |
DESIGN |
a design matrix to indicate if rows belong to groups. |
make_design_nominal |
a boolean. If TRUE (default), DESIGN is a vector that indicates groups (and will be dummy-coded). If FALSE, DESIGN is a dummy-coded matrix. |
weights1 |
a diagonal matrix or column-vector of weights for the columns of DATA1 |
weights2 |
a diagonal matrix or column-vector of weights for the columns of DATA2 |
symmetric |
a boolean. If TRUE (default) symmetric factor scores for rows. |
graphs |
a boolean. If TRUE (default), graphs and plots are provided (via |
k |
number of components to return. |
This implementation of Partial Least Squares is for two categorical data sets (Beaton et al., 2013), and based on the PLS method proposed by Tucker (1958) and again by Bookstein (1994).
See epCA
(and also coreCA
) for details on what is returned. In addition to the values returned:
W1 |
Weights for columns of DATA1, replaces |
W2 |
Weights for columns of DATA2, replaces |
lx |
latent variables from DATA1 computed for observations |
ly |
latent variables from DATA2 computed for observations |
Derek Beaton, Hervé Abdi
Tucker, L. R. (1958). An inter-battery method of factor analysis. Psychometrika, 23(2), 111–136.
Bookstein, F., (1994). Partial least squares: a dose–response model for measurement in the behavioral and brain sciences. Psycoloquy 5 (23)
Abdi, H. (2007). Singular Value Decomposition (SVD) and Generalized Singular Value Decomposition (GSVD). In N.J. Salkind (Ed.): Encyclopedia of Measurement and Statistics.Thousand Oaks (CA): Sage. pp. 907-912.
Krishnan, A., Williams, L. J., McIntosh, A. R., & Abdi, H. (2011). Partial Least Squares (PLS) methods for neuroimaging: A tutorial and review. NeuroImage, 56(2), 455 – 475.
Beaton, D., Filbey, F., & Abdi H. (in press, 2013). Integrating partial least squares correlation and correspondence analysis for nominal data. In Abdi, H., Chin, W., Esposito Vinzi, V., Russolillo, G., & Trinchera, L. (Eds.), New Perspectives in Partial Least Squares and Related Methods. New York: Springer Verlag.
data(snps.druguse) plsca.res <- tepPLSCA(snps.druguse$DATA1,snps.druguse$DATA2, make_data1_nominal=TRUE,make_data2_nominal=TRUE)
data(snps.druguse) plsca.res <- tepPLSCA(snps.druguse$DATA1,snps.druguse$DATA2, make_data1_nominal=TRUE,make_data2_nominal=TRUE)
TExPosition's DESIGN matrix check function. Calls into ExPosition's designCheck
.
texpoDesignCheck(DATA = NULL, DESIGN = NULL, make_design_nominal = TRUE, force_bary=FALSE)
texpoDesignCheck(DATA = NULL, DESIGN = NULL, make_design_nominal = TRUE, force_bary=FALSE)
DATA |
original data that should be matched to a design matrix |
DESIGN |
a column vector with levels for observations or a dummy-coded matrix |
make_design_nominal |
a boolean. Will make DESIGN nominal if TRUE (default). |
force_bary |
a boolean. If TRUE, it forces the check for barycentric methods (tepDICA, tepBADA). If FALSE, |
For BADA & DICA, execution stops if:
1. DESIGN has more columns (groups) than observations,
2. DESIGN has only 1 column (group), or
3. DESIGN has at least 1 occurence where an observation is the only observation in a group (i.e., colSums(DESIGN)==1 at least once).
DESIGN |
dummy-coded design matrix |
Derek Beaton