SelvarLearnLasso: Regularization for variable selection in discriminant...

In masedki/SelvarMix: Regularization for variable selection in model-based clustering and discriminant analysis.

Description

This function implements the variable selection in discriminant analysis using a lasso ranking on the variables as described in Sedki et al (2014). The variable ranking step uses the penalized EM algorithm of Zhou et al (2009) (adapted in Sedki et al (2014) for the discriminant analysis settings). A testing sample can be used to compute the averaged classification error rate.

Usage

SelvarLearnLasso(data, knownlabels, lambda, rho, hybrid.size,  models, 
                 regModel, indepModel, dataTest, labelsTest, nbCores)

Arguments

data	matrix containing quantitative data. Rows correspond to observations and columns correspond to variables
knownlabels	an integer vector or a factor of size number of observations. Each cell corresponds to a cluster affectation
lambda	numeric listing of tuning parameter for \ell_1 mean penalty
rho	numeric listing of tuning parameter for \ell_1 precision matrix penalty
hybrid.size	optional parameter make less strength the hybrid forward and backward algorithms to select S and W sets

models	a Rmixmod [Model] object defining the list of models to run. The models Gaussian_pk_L_C, Gaussian_pk_Lk_C, Gaussian_pk_L_Ck, and Gaussian_pk_Lk_Ck are called by default (see mixmodGaussianModel() in Rmixmod package to specify other models)
regModel	list of character defining the covariance matrix form for the linear regression of U on the R set of variable. Possible values: "LI" for spherical form, "LB" for diagonal form and "LC" for general form. Possible values: "LI", "LB", "LC", c("LI", "LB") , c("LI", "LC"), c("LB", "LC") and c("LI", "LB", "LC"). Default is c("LI", "LB", "LC")
indepModel	list of character defining the covariance matrix form for independent variables W. Possible values: "LI" for spherical form and "LB" for diagonal form. Possible values: "LI", "LB", c("LI", "LB"). Default is c("LI", LB")
dataTest	matrix containing quantitative testing data. Rows correspond to observations and columns correspond to variables
labelsTest	an integer vector or a factor of size number of testing observations. Each cell corresponds to a cluster affectation
nbCores	number of CPUs to be used when parallel computing is utilized (default is 2)

Value

S	The selected set of relevant clustering variables
R	The selected subset of regressors
U	The selected set of redundant variables
W	The selected set of independent variables
criterionValue	The criterion value for the selected model
nbCluster	The selected number of clusters
model	The selected covariance model

regModel	The selected covariance form for the regression
indepModel	The selected covariance form for the independent variables
proba	Optional : matrix containing the conditional probabilities of belonging to each cluster for the testing observations
partition	Optional: vector containing the cluster assignments of the testing observations according to the Maximum-a-Posteriori rule
error	Optional : error rate done by the predicted partition (obtained using Maximum-A-Posteriori rule)

Author(s)

Mohammed Sedki <[e-mail]>

References

Zhou, H., Pan, W., and Shen, X., 2009. "Penalized model-based clustering with unconstrained covariance matrices". Electronic Journal of Statistics, vol. 3, pp.1473-1496.

Maugis, C., Celeux, G., and Martin-Magniette, M. L., 2009. "Variable selection in model-based clustering: A general variable role modeling". Computational Statistics and Data Analysis, vol. 53/11, pp. 3872-3882.

Sedki, M., Celeux, G., Maugis-Rabusseau, C., 2014. "SelvarMix: A R package for variable selection in model-based clustering and discriminant analysis with a regularization approach". Inria Research Report available at http://hal.inria.fr/hal-01053784

See Also

SelvarClustLasso SortvarLearn SortvarClust scenarioCor

Examples

## Not run: 
## Simulated data  example as shown in Sedki et al (2014)
require(Rmixmod)
require(glasso)
data(scenarioCor)

lambda <- seq(20,  50, length = 10)
rho <- seq(1, 2, length=2)
hybrid.size <- 3
models <- mixmodGaussianModel(family = "spherical", equal.proportions = TRUE)
regModel <- c("LI","LB","LC")
indepModel <- c("LI","LB")

## variables selection in discriminant analysis
## training sample : n = 1900 observations , p = 14 variables 
data.learn <- scenarioCor[1:1900,1:14]
labels.learn <-scenarioCor[1:1900,15]

## testing sample : n = 100 observations, p = 14 variables
data.test <- scenarioCor[1901:2000,1:14]
labels.test <-scenarioCor[1901:2000,15]

simulate.da <- SelvarLearnLasso(data.learn, labels.learn, lambda, rho, hybrid.size, 
                                models, regModel, indepModel, data.test, labels.test)

## End(Not run)

masedki/SelvarMix documentation built on May 21, 2019, 12:42 p.m.