SelvarMix is a package where a regularization approach of variable selection is considered in model-based clustering and discriminant analysis frameworks. First, this procedure consists of ranking the variables with a lasso-like procedure. Second, the method of Maugis et al (2009, 2011) is adapted to define the role of variables in the two frameworks. SelvarMix provides a faster variable selection algorithm than the backward stepwise or forward stepwise algorithms of Maugis et al (2009), allowing us to study high-dimensional datasets.
|License:||GPL-3 + file LICENSE|
The general purpose of the package is to perform variable selection in model-based clustering and discriminant analysis. It focus on model-based clustering, where the clusters are assumed to arise from Gaussian distributions. The most achieved model in model-based clustering has been proposed by Maugis et al (2009). This so-called SRUW modeling considers three roles of variables: one variable may belong to the relevant clustering set S, the redundant variable set U or the independent variable set W. Moreover, the redundant variables may be explained by a subset R of the relevant variables S. In order to avoid the greedy algorithms when high-dimensional data are studied, the SelvarMix procedure is proposed. It proceeds in two steps: First, the variables are ranked using a lasso-like procedure analogous to the one of Zhou et al (2009); second, the SRUW procedure is run on this ranked set of variables.
Author: Mohammed Sedki, Gilles Celeux and Cathy Maugis-Rabusseau
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.
Maugis, C., Celeux, G., and Martin-Magniette, M. L., 2011. "Variable selection in model-based discriminant analysis". Journal of Multivariate Analysis, vol. 102, pp. 1374-1387.
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.
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
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
## Not run: ## wine data set ## n = 178 observations, p = 27 variables data(wine) ## variable selection in model-based clustering set.seed(123) obj <- SelvarClustLasso(x=wine[,1:27], nbcluster=1:5, nbcores=4) summary(obj) print(obj) ## variables selection in discriminant analysis set.seed(123) a <- seq(1, 178, 10) b <- setdiff(1:178, a) obj <- SelvarLearnLasso(x=wine[b,1:27], z=wine[b,28], xt=wine[a,1:27], zt=wine[a,28], nbcores=4) summary(obj) print(obj) ## End(Not run)