|
Goto
|
Name
|
Description
|
Status
|
Version
|
Update
|
|
activeset
|
Primal Active Set Method for QP. Will be merged with dykstra and esopo. Only for restricted least squares.
|
bare bones.
|
0.1.00
|
02/24/07
|
|
|
alscal
|
Metric MDS on squared distances. Minimizes sstress using cyclic cooordinate descend by solving cubic
polynomial equations. More or less superseded by elegant. Available for
symmetric and rectangular matrices.
|
partly done.
|
0.1.2
|
02/07/06
|
|
aspect
|
Optimizing functions of correlation
matrices over transformations/quantifications of the variables --
functions implemented are sums of eigenvalues, determinants, multiple
correlations, power sums. Also includes lineals function to optimally
linearize all bivariate regressions.
|
done.
|
0.1.1
|
04/15/06
|
|
astro
|
Techniques for fitting an ellipse to a set of points in the plane
or in space (mostly for applications in astronomy).
|
first pass.
|
0.0.01
|
05/22/07
|
|
ca
|
Correspondence analysis and canonical correspondence analysis with many
plotting options (biplots, row and column plots, benzecri plots, transformation
plots, graph plots, regression plots, ...). This also has Nora's algorithm
for CA with missing data.
|
done.
|
1.5
|
01/15/06
|
|
distassoc
|
Fits the distance association model maximizing the Poisson likelihood using
majorization and MDS. This is a exponential distance version of CA. Includes
Ideal Point Discriminant Analysis. Can fit both exponents of distances and
exponents of squared distances to either square or rectangular matrices. Allows
for linear constraints. A version specifically designed for indicator matrices,
which fits the same representation as the logitfold package, is also included.
Also does quasi-symmetry and quasi-independence.
|
partly done.
|
0.4.06
|
03/07/06
|
|
derivatives
|
Generalized eigenvalue and generalized singular value algorithms for matrices that are
functions of parameters. Also computes the partial derivatives of generalized eigen/singular values
and generalized eigen/singular vectors wrt the parameters.
|
first also
|
0.0.1
|
12/15/07
|
|
dykstra
|
Convex programming by Dykstra/Bregman dual projection methods.
basic Dykstra. alternating projections for cone and subspace.
|
partly done
|
0.2.00
|
02/24/07
|
|
elegant
|
Metric MDS on squared distances minimizing sstress. Uses the majorization method.
Currently for square symmetric and rectangular matrices. Different majorization
bounds can be chosen.This version is n/2 times as fast as the original "elegant"
algorithm, where n is the number of points.
|
partly done.
|
0.2.2
|
02/22/06
|
|
esopo
|
Quadratic programming by cyclic coordinate ascend. Applications to monotone regression.
Also fits the best metric space to a finite set of points (i.e. finds best fitting
numbers satisfying all triangle inequalities). Also has a CCA algorithm for fitting
monotone polynomials and splines.
|
mostly done.
|
0.2.00
|
02/17/07
|
|
gifi
|
Update to homals on CRAN. Code for
multiple correspondence analysis, nonlinear principal
component analysis, nonlinear multiset canonical
correlation analysis. Implements everything in Gifi (1990)
or SPSS Categories, and then some. Many different plots,
such as Voronoi plots. Quicktime movies of iterations.
|
mostly done.
|
2.0.0
|
05/04/06
|
|
ipscal
|
Metric MDS on the inner products using strain. One matrix
only.
|
stagnant.
|
0.0.1
|
|
|
irt
|
Routines for IRT.
|
planning stage.
|
|
|
|
logitfold
|
Logistic likelihood solution to nonmetric
unfolding for binary and multicategory data, multidimensional
choice models, logistic principal component analysis and
canonical analysis, roll-call models, multidimensional Rasch models.
This is the exponential or likelihood version of MCA (Gifi Goes
Logistic). Distance version.
|
partly done.
|
|
|
|
logreg
|
Majorization methods for logistic regression. Newton, quadratic,
cubic, quartic majorization.
|
basics done.
|
1.0.0
|
03/27/06
|
|
logithom
|
Logistic likelihood solution to nonmetric
unfolding for binary and multicategory data, multidimensional
choice models, logistic principal component analysis and
canonical analysis, roll-call models, multidimensional Rasch models.
This is the exponential or likelihood version of MCA (Gifi Goes
Logistic). Inner product version.
|
basics done.
|
1.4.0
|
05/12/06
|
|
lowpoly
|
Solve one-dimensional polynomial equations of orders 1-4 for
all the real roots. Also has a routine for finding the real
minimizer of a quartic.
|
done.
|
|
|
|
lowrank
|
Low-rank approximation of matrices. LS and LAV. Huber loss. Symmetric and
rectangular. Missing data. Majorization.
|
started.
|
0.3.0
|
04/16/06
|
|
lssem
|
Linear structural equation models with latent
variables, reformulated as matrix decomposition problems, and
solved by majorizing least squares loss functions. Article
included. Also article and code for least squares factor analysis.
|
factor analysis code.
|
0.1.0
|
10/21/06
|
|
mdsutils
|
Some multidimensional scaling utilities used by other packages.
|
done for now
|
0.2.0
|
05/24/07
|
|
|
multilevel
|
Programs for multilevel analysis.
|
almost nothing.
|
|
|
|
multiway
|
Multiway generalizations of principal component
analysis with various constraints on the decomposition.
|
on the horizon.
|
|
|
|
|
nestorov
|
Routines for logistic regression using different quadratic and
cubic majorization methods. Also has a safe-guarded Newton's
method (which is part majorization, part Newton).
|
done.
|
1.0.0
|
03/27/06
|
|
|
nextperm
|
Routines for least squares undimensional scaling by testing all permutations
in lexicographical order. This does not really work for n>12, but
I have some ideas to improve it.
|
done.
|
|
|
|
|
nlpca
|
Miscellaneous nonlinear principal component analysis techniques. Currently
has PCA for rank-orders (using least squares optimal scaling). Also text of
CARME chapter. Also algorithm for fitting monotone polynomials and splines.
|
0.1.00.
|
2007-02-24
|
|
|
|
pava
|
Pooled Adjacent Violators algorithm for general monotone regression.
Has subroutines for weighted median, weighted mean, and weighted
fractiles. Also routines for primary, secondary, and tertiary approach
to ties.
|
done.
|
|
|
|
probit
|
The probit version of logitfold. Does probit regression, fitting
of discreticized multivariate normal, probit PCA, probit IRT, and probit factor
analysis. The probit version of MCA. Gifi joins Pearson. Will be
extended to cover most of SEM.
|
basically done.
|
0.7.1
|
03/31/06
|
|
scalassoc
|
Scalar product version of distassoc, applied to cross tables, and interaction tables.
Remember: the exponential scalar product model with bias is equivalent
to the exponential squared distance model with bias.
|
basically done.
|
1.0.0
|
5/05/06
|
|
singlepeaked
|
Exponential distance method for single peaked choices.
|
paper and first alpha
|
0.0.1
|
05/12/06
|
|
smacofQuadratic
|
Metric and nonmetric multidimensional scaling methods for putting points on spheres, ellipses, and other
quadratic surfaces, using either Euclidean distance or geodetic distance.
|
first pass.
|
0.1.0
|
05/24/07
|
|
smacofPlain
|
Metric and nonmetric multidimensional
scaling minimizing stress, for symmetric and rectangular matrices.
|
mostly done.
|
0.6.00
|
02/22/06
|
|
smacofConstrained
|
Metric and nonmetric multidimensional
scaling minimizing stress, with simple constraints on the configuration.
|
mostly done.
|
0.6.00
|
02/22/06
|
|
smacofThreeWay
|
Metric and nonmetric multidimensional
scaling minimizing stress, using individual difference models on three-way matrices.
|
mostly done.
|
0.6.00
|
02/22/06
|
|
voronoi
|
Pseudo-voronoi diagrams for exponential inner product and squared distance
models with and without bias. Only the squared distance model without bias
actually gives a true Voronoi diagram.
|
mostly done.
|
1.0.1
|
04/26/06
|