similar to: softImpute_1.0 uploaded to CRAN

We have put a new package sparsenet on CRAN. Sparsenet fits regularization paths for sparse model selection via coordinate descent, using a penalized least-squares framework and a non-convex penalty. The package is based on our JASA paper Rahul Mazumder, Jerome Friedman and Trevor Hastie: SparseNet : Coordinate Descent with Non-Convex Penalties. (JASA 2011)

We have put a new package sparsenet on CRAN. Sparsenet fits regularization paths for sparse model selection via coordinate descent, using a penalized least-squares framework and a non-convex penalty. The package is based on our JASA paper Rahul Mazumder, Jerome Friedman and Trevor Hastie: SparseNet : Coordinate Descent with Non-Convex Penalties. (JASA 2011)

This update adds an intercept option (by popular request) - now one can fit a model without an intercept Glmnet is a package that fits the regularization path for a number of generalized linear models, with with "elastic net" regularization (tunable mixture of L1 and L2 penalties). Glmnet uses pathwise coordinate descent, and is very fast. The current list of models covered are:

This update adds an intercept option (by popular request) - now one can fit a model without an intercept Glmnet is a package that fits the regularization path for a number of generalized linear models, with with "elastic net" regularization (tunable mixture of L1 and L2 penalties). Glmnet uses pathwise coordinate descent, and is very fast. The current list of models covered are:

This is a new version of glmnet, that incorporates some bug fixes and speedups. * a new convergence criterion which which offers 10x or more speedups for saturated fits (mainly effects logistic, Poisson and Cox) * one can now predict directly from a cv.object - see the help files for cv.glmnet and predict.cv.glmnet * other new methods are deviance() for "glmnet" and coef() for

Dear R-list, Among the (R-)tools, I've seen on the net, for (bivariate) Principal Component scatter plots (+histograms), "plotpc" [1] is the one I like most. By default it performs PCA on a bivariate dataset based on R's "princomp()" (which is the eigenvector-based algebraic solution to PCA). I would like to modify "plotpc()" in order be able, as an

glmnet is a package that fits the regularization path for linear, two- and multi-class logistic regression models with "elastic net" regularization (tunable mixture of L1 and L2 penalties). glmnet uses pathwise coordinate descent, and is very fast. Some of the features of glmnet: * by default it computes the path at 100 uniformly spaced (on the log scale) values of the

glmnet is a package that fits the regularization path for linear, two- and multi-class logistic regression models with "elastic net" regularization (tunable mixture of L1 and L2 penalties). glmnet uses pathwise coordinate descent, and is very fast. Some of the features of glmnet: * by default it computes the path at 100 uniformly spaced (on the log scale) values of the

glmnet_1.2 has been uploaded to CRAN. This is a major upgrade, with the following additional features: * poisson family, with dense or sparse x * Cox proportional hazards family, for dense x * wide range of cross-validation features. All models have several criteria for cross-validation. These include deviance, mean absolute error, misclassification error and "auc" for logistic or

glmnet_1.2 has been uploaded to CRAN. This is a major upgrade, with the following additional features: * poisson family, with dense or sparse x * Cox proportional hazards family, for dense x * wide range of cross-validation features. All models have several criteria for cross-validation. These include deviance, mean absolute error, misclassification error and "auc" for logistic or

We have uploaded to CRAN the first version of glmpath, which fits the L1 regularization path for generalized linear models. The lars package fits the entire piecewise-linear L1 regularization path for the lasso. The coefficient paths for L1 regularized glms, however, are not piecewise linear. glmpath uses convex optimization - in particular predictor-corrector methods- to fit the

We have uploaded to CRAN the first version of glmpath, which fits the L1 regularization path for generalized linear models. The lars package fits the entire piecewise-linear L1 regularization path for the lasso. The coefficient paths for L1 regularized glms, however, are not piecewise linear. glmpath uses convex optimization - in particular predictor-corrector methods- to fit the

Lately I'm getting this error quite a bit: Error in La.svd(x, nu, nv) : error code 1 from Lapack routine 'dgesdd' I'm running R 2.5.0 on a 64 bit Intel machine running Fedora (8 I think). Maybe the 64 bit platform is more fragile about declaring convergence. I'm seeing way more of these errors than I ever have before. From R-Help I see that this issue comes up from time to

After fixing princomp(), recently, {tiny negative eigen-values are possible for non-negative definite matrices} Fritz Leisch drew my attention to the fact the not only eigen() can be funny, but also svd(). Adrian Trappleti found that the singular values returned can be "-0" instead of "0". This will be a problem in something like sd <- svd(Mat) $ d

Dear R-devel: I could use some advice about matrix calculations and steps that might make for faster computation of generalized inverses. It appears in some projects there is a bottleneck at the use of svd in calculation of generalized inverses. Here's some Rprof output I need to understand. > summaryRprof("Amelia.out") $by.self self.time self.pct

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi all, this is probably simple and I'm just doing something stupid, sorry about that :-) I'm trying to convert words (strings of letters) into a fairly small dimensional space (say 10, but anything between about 5 and 50 would be ok), which I will call a feature vector. The the distance between two words represents the similarity of the

We have uploaded to CRAN a new version of glmpath, a package which fits the L1 regularization path for generalized linear models. The revision includes: - coxpath, a function for fitting the L1-regularization path for the Cox ph model; - bootstrap functions for analyzing sparse solutions; - the ability to mix in L2 regularization along with L1 (elasticnet). We have also completed a report that

We have uploaded to CRAN a new version of glmpath, a package which fits the L1 regularization path for generalized linear models. The revision includes: - coxpath, a function for fitting the L1-regularization path for the Cox ph model; - bootstrap functions for analyzing sparse solutions; - the ability to mix in L2 regularization along with L1 (elasticnet). We have also completed a report that

Hi, I'm trying to do PCA on a n by p wide matrix (n < p), and I'd like to get more principal components than there are rows. However, svd() only returns a V matrix of with n columns (instead of p) unless the argument nv=p is set (prcomp calls svd without setting it). Moreover, the eigenvalues returned are always min(n, p) instead of p, even if nv is set: > x <-

Prof. Bates & R-devel, I've done more test with the following results: I have two versions of ATLAS 3.2.1. One was compiled on my old Thinkpad 600E (PII), the other was compiled on my new Thinkpad T22 (PIIISSE1). I compiled R-devel dated 10/31, 11/01 and 11/04, linked against either of the two ATLAS libs. All gave Dr. Watson when given this code: La.svd(matrix(runif(1e5), 1e3,