tag:crantastic.org,2005:/authors/8834Latest activity for Haoming Jiang2019-10-28T15:21:42Zcrantastic.orgtag:crantastic.org,2005:TimelineEvent/941552019-10-28T15:21:42Z2019-10-28T15:21:42Zhuge was upgraded to version 1.3.4<a href="/packages/huge">huge</a> was <span class="action">upgraded</span> to version <a href="/packages/huge/versions/89514">1.3.4</a><br /><h3>Package description:</h3><p>Provides a general framework for high-dimensional undirected graph estimation. It integrates data preprocessing, neighborhood screening, graph estimation, and model selection techniques into a pipeline. In preprocessing stage, the nonparanormal(npn) transformation is applied to help relax the normality assumption. In the graph estimation stage, the graph structure is estimated by Meinshausen-Buhlmann graph estimation or the graphical lasso, and both methods can be further accelerated by the lossy screening rule preselecting the neighborhood of each variable by correlation thresholding. We target on high-dimensional data analysis usually d >> n, and the computation is memory-optimized using the sparse matrix output. We also provide a computationally efficient approach, correlation thresholding graph estimation. Three regularization/thresholding parameter selection methods are included in this package: (1)stability approach for regularization selection (2) rotation information criterion (3) extended Bayesian information criterion which is only available for the graphical lasso.</p>crantastic.orgtag:crantastic.org,2005:TimelineEvent/925412019-09-09T21:21:51Z2019-09-09T21:21:51Zhuge was upgraded to version 1.3.3<a href="/packages/huge">huge</a> was <span class="action">upgraded</span> to version <a href="/packages/huge/versions/87999">1.3.3</a><br /><h3>Package description:</h3><p>Provides a general framework for high-dimensional undirected graph estimation. It integrates data preprocessing, neighborhood screening, graph estimation, and model selection techniques into a pipeline. In preprocessing stage, the nonparanormal(npn) transformation is applied to help relax the normality assumption. In the graph estimation stage, the graph structure is estimated by Meinshausen-Buhlmann graph estimation or the graphical lasso, and both methods can be further accelerated by the lossy screening rule preselecting the neighborhood of each variable by correlation thresholding. We target on high-dimensional data analysis usually d >> n, and the computation is memory-optimized using the sparse matrix output. We also provide a computationally efficient approach, correlation thresholding graph estimation. Three regularization/thresholding parameter selection methods are included in this package: (1)stability approach for regularization selection (2) rotation information criterion (3) extended Bayesian information criterion which is only available for the graphical lasso.</p>crantastic.orgtag:crantastic.org,2005:TimelineEvent/901452019-07-02T07:43:04Z2019-07-02T07:43:04ZSAM was upgraded to version 1.1.2.1<a href="/packages/SAM">SAM</a> was <span class="action">upgraded</span> to version <a href="/packages/SAM/versions/85734">1.1.2.1</a><br /><h3>Package description:</h3><p>Computationally efficient tools for high dimensional predictive modeling (regression and classification). SAM is short for sparse additive modeling, and adopts the computationally efficient basis spline technique. We solve the optimization problems by various computational algorithms including the block coordinate descent algorithm, fast iterative soft-thresholding algorithm, and newton method. The computation is further accelerated by warm-start and active-set tricks.</p>crantastic.orgtag:crantastic.org,2005:TimelineEvent/870252019-04-08T12:21:37Z2019-04-08T12:21:37Zhuge was upgraded to version 1.3.2<a href="/packages/huge">huge</a> was <span class="action">upgraded</span> to version <a href="/packages/huge/versions/82726">1.3.2</a><br /><h3>Package description:</h3><p>Provides a general framework for high-dimensional undirected graph estimation. It integrates data preprocessing, neighborhood screening, graph estimation, and model selection techniques into a pipeline. In preprocessing stage, the nonparanormal(npn) transformation is applied to help relax the normality assumption. In the graph estimation stage, the graph structure is estimated by Meinshausen-Buhlmann graph estimation or the graphical lasso, and both methods can be further accelerated by the lossy screening rule preselecting the neighborhood of each variable by correlation thresholding. We target on high-dimensional data analysis usually d >> n, and the computation is memory-optimized using the sparse matrix output. We also provide a computationally efficient approach, correlation thresholding graph estimation. Three regularization/thresholding parameter selection methods are included in this package: (1)stability approach for regularization selection (2) rotation information criterion (3) extended Bayesian information criterion which is only available for the graphical lasso.</p>crantastic.orgtag:crantastic.org,2005:TimelineEvent/869962019-04-07T18:42:46Z2019-04-07T18:42:46ZSAM was upgraded to version 1.1.2<a href="/packages/SAM">SAM</a> was <span class="action">upgraded</span> to version <a href="/packages/SAM/versions/82697">1.1.2</a><br /><h3>Package description:</h3><p>Computationally efficient tools for high dimensional predictive modeling (regression and classification). SAM is short for sparse additive modeling, and adopts the computationally efficient basis spline technique. We solve the optimization problems by various computational algorithms including the block coordinate descent algorithm, fast iterative soft-thresholding algorithm, and newton method. The computation is further accelerated by warm-start and active-set tricks.</p>crantastic.orgtag:crantastic.org,2005:TimelineEvent/858232019-03-12T00:02:53Z2019-03-12T00:02:53ZSAM was upgraded to version 1.1.1<a href="/packages/SAM">SAM</a> was <span class="action">upgraded</span> to version <a href="/packages/SAM/versions/81608">1.1.1</a><br /><h3>Package description:</h3><p>Computationally efficient tools for high dimensional predictive modeling (regression and classification). SAM is short for sparse additive modeling, and adopts the computationally efficient basis spline technique. We solve the optimization problems by various computational algorithms including the block coordinate descent algorithm, fast iterative soft-thresholding algorithm, and newton method. The computation is further accelerated by warm-start and active-set tricks.</p>crantastic.orgtag:crantastic.org,2005:TimelineEvent/858212019-03-11T23:21:37Z2019-03-11T23:21:37Zhuge was upgraded to version 1.3.1<a href="/packages/huge">huge</a> was <span class="action">upgraded</span> to version <a href="/packages/huge/versions/81606">1.3.1</a><br /><h3>Package description:</h3><p>Provides a general framework for high-dimensional undirected graph estimation. It integrates data preprocessing, neighborhood screening, graph estimation, and model selection techniques into a pipeline. In preprocessing stage, the nonparanormal(npn) transformation is applied to help relax the normality assumption. In the graph estimation stage, the graph structure is estimated by Meinshausen-Buhlmann graph estimation or the graphical lasso, and both methods can be further accelerated by the lossy screening rule preselecting the neighborhood of each variable by correlation thresholding. We target on high-dimensional data analysis usually d >> n, and the computation is memory-optimized using the sparse matrix output. We also provide a computationally efficient approach, correlation thresholding graph estimation. Three regularization/thresholding parameter selection methods are included in this package: (1)stability approach for regularization selection (2) rotation information criterion (3) extended Bayesian information criterion which is only available for the graphical lasso.</p>crantastic.orgtag:crantastic.org,2005:TimelineEvent/850442019-02-22T07:21:39Z2019-02-22T07:21:39Zhuge was upgraded to version 1.3.0<a href="/packages/huge">huge</a> was <span class="action">upgraded</span> to version <a href="/packages/huge/versions/80851">1.3.0</a><br /><h3>Package description:</h3><p>Provides a general framework for high-dimensional undirected graph estimation. It integrates data preprocessing, neighborhood screening, graph estimation, and model selection techniques into a pipeline. In preprocessing stage, the nonparanormal(npn) transformation is applied to help relax the normality assumption. In the graph estimation stage, the graph structure is estimated by Meinshausen-Buhlmann graph estimation or the graphical lasso, and both methods can be further accelerated by the lossy screening rule preselecting the neighborhood of each variable by correlation thresholding. We target on high-dimensional data analysis usually d >> n, and the computation is memory-optimized using the sparse matrix output. We also provide a computationally efficient approach, correlation thresholding graph estimation. Three regularization/thresholding parameter selection methods are included in this package: (1)stability approach for regularization selection (2) rotation information criterion (3) extended Bayesian information criterion which is only available for the graphical lasso.</p>crantastic.orgtag:crantastic.org,2005:TimelineEvent/849132019-02-18T16:22:53Z2019-02-18T16:22:53ZSAM was upgraded to version 1.1.0<a href="/packages/SAM">SAM</a> was <span class="action">upgraded</span> to version <a href="/packages/SAM/versions/80724">1.1.0</a><br /><h3>Package description:</h3><p>Computationally efficient tools for high dimensional predictive modeling (regression and classification). SAM is short for sparse additive modeling, and adopts the computationally efficient basis spline technique. We solve the optimization problems by various computational algorithms including the block coordinate descent algorithm, fast iterative soft-thresholding algorithm, and newton method. The computation is further accelerated by warm-start and active-set tricks.</p>crantastic.orgtag:crantastic.org,2005:TimelineEvent/494562016-03-12T00:15:36Z2016-03-12T00:15:36Zcrantastic_production tagged huge with gR<a href="/users/146">crantastic_production</a> <span class="action">tagged</span> <a href="/packages/huge">huge</a> with <a href="/task_views/gR">gR</a>crantastic_productiontag:crantastic.org,2005:TimelineEvent/436982015-09-16T08:11:14Z2015-09-16T08:11:14Zhuge was upgraded to version 1.2.7<a href="/packages/huge">huge</a> was <span class="action">upgraded</span> to version <a href="/packages/huge/versions/43490">1.2.7</a><br /><h3>Package description:</h3><p>Provides a general framework for high-dimensional undirected graph estimation. It integrates data preprocessing, neighborhood screening, graph estimation, and model selection techniques into a pipeline. In preprocessing stage, the nonparanormal(npn) transformation is applied to help relax the normality assumption. In the graph estimation stage, the graph structure is estimated by Meinshausen-Buhlmann graph estimation or the graphical lasso, and both methods can be further accelerated by the lossy screening rule preselecting the neighborhood of each variable by correlation thresholding. We target on high-dimensional data analysis usually d >> n, and the computation is memory-optimized using the sparse matrix output. We also provide a computationally efficient approach, correlation thresholding graph estimation. Three regularization/thresholding parameter selection methods are included in this package: (1)stability approach for regularization selection (2) rotation information criterion (3) extended Bayesian information criterion which is only available for the graphical lasso.</p>crantastic.orgtag:crantastic.org,2005:TimelineEvent/312752013-04-01T22:51:06Z2013-04-01T22:51:06ZSAM was upgraded to version 1.0.3<a href="/packages/SAM">SAM</a> was <span class="action">upgraded</span> to version <a href="/packages/SAM/versions/26532">1.0.3</a><br /><h3>Package description:</h3><p>The package SAM targets at high dimensional predictive modeling (regression and classification) for complex data analysis. SAM is short for sparse additive modeling, and adopts the computationally efficient basis spline technique. We solve the optimization problems by various computational algorithms including the block coordinate descent algorithm, fast iterative soft-thresholding algorithm, and newton method. The computation is further accelerated by warm-start and active-set tricks.</p>crantastic.orgtag:crantastic.org,2005:TimelineEvent/261492012-10-15T16:35:34Z2012-10-15T16:35:34ZSAM was released<a href="/packages/SAM">SAM</a> was <span class="action">released</span><br /><h3>Package description:</h3><p>Computationally efficient tools for high dimensional predictive modeling (regression and classification). SAM is short for sparse additive modeling, and adopts the computationally efficient basis spline technique. We solve the optimization problems by various computational algorithms including the block coordinate descent algorithm, fast iterative soft-thresholding algorithm, and newton method. The computation is further accelerated by warm-start and active-set tricks.</p>crantastic.orgtag:crantastic.org,2005:TimelineEvent/245862012-08-16T06:11:00Z2012-08-16T06:11:00Zhuge was upgraded to version 1.2.4<a href="/packages/huge">huge</a> was <span class="action">upgraded</span> to version <a href="/packages/huge/versions/20598">1.2.4</a><br /><h3>Package description:</h3><p>The package "huge" provides a general framework for high-dimensional undirected graph estimation. It integrates data preprocessing, neighborhood screening, graph estimation, and model selection techniques into a pipeline. In preprocessing stage, the nonparanormal(npn) transformation is applied to help relax the normality assumption. In the graph estimation stage, the graph structure is estimated by Meinshausen-Buhlmann graph estimation or the graphical lasso, and both methods can be further accelerated by the lossy screening rule preselecting the neighborhood of each variable by correlation thresholding. We target on high-dimensional data analysis usually d >> n, and the computation is memory-optimized using the sparse matrix output. We also provide a computationally efficient approach, correlation thresholding graph estimation. Three regularization/thresholding parameter selection methods are included in this package: (1)stability approach for regularization selection (2) rotation information criterion (3) extended Bayesian information criterion which is only available for the graphical lasso.</p>crantastic.orgtag:crantastic.org,2005:TimelineEvent/206582012-03-22T08:30:34Z2012-03-22T08:30:34Zhuge was upgraded to version 1.2.3<a href="/packages/huge">huge</a> was <span class="action">upgraded</span> to version <a href="/packages/huge/versions/17721">1.2.3</a><br /><h3>Package description:</h3><p>The package "huge" provides a general framework for high-dimensional undirected graph estimation. It integrates data preprocessing, neighborhood screening, graph estimation, and model selection techniques into a pipeline. In preprocessing stage, the nonparanormal(npn) transformation is applied to help relax the normality assumption. In the graph estimation stage, the graph structure is estimated by Meinshausen-Buhlmann graph estimation or the graphical lasso, and both methods can be further accelerated by the lossy screening rule preselecting the neighborhood of each variable by correlation thresholding. We target on high-dimensional data analysis usually d >> n, and the computation is memory-optimized using the sparse matrix output. We also provide a computationally efficient approach, correlation thresholding graph estimation. Three regularization/thresholding parameter selection methods are included in this package: (1)stability approach for regularization selection (2) rotation information criterion (3) extended Bayesian information criterion which is only available for the graphical lasso.</p>crantastic.orgtag:crantastic.org,2005:TimelineEvent/206112012-03-21T08:10:39Z2012-03-21T08:10:39Zhuge was upgraded to version 1.2.2<a href="/packages/huge">huge</a> was <span class="action">upgraded</span> to version <a href="/packages/huge/versions/17683">1.2.2</a><br /><h3>Package description:</h3><p>The package "huge" provides a general framework for high-dimensional undirected graph estimation. It integrates data preprocessing, neighborhood screening, graph estimation, and model selection techniques into a pipeline. In preprocessing stage, the nonparanormal(npn) transformation is applied to help relax the normality assumption. In the graph estimation stage, the graph structure is estimated by Meinshausen-Buhlmann graph estimation or the graphical lasso, and both methods can be further accelerated by the lossy screening rule preselecting the neighborhood of each variable by correlation thresholding. We target on high-dimensional data analysis usually d >> n, and the computation is memory-optimized using the sparse matrix output. We also provide a computationally efficient approach, correlation thresholding graph estimation. Three regularization/thresholding parameter selection methods are included in this package: (1)stability approach for regularization selection (2) rotation information criterion (3) extended Bayesian information criterion which is only available for the graphical lasso.</p>crantastic.orgtag:crantastic.org,2005:TimelineEvent/155592011-08-22T20:10:28Z2011-08-22T20:10:28Zhuge was upgraded to version 1.1.2<a href="/packages/huge">huge</a> was <span class="action">upgraded</span> to version <a href="/packages/huge/versions/13889">1.1.2</a><br /><h3>Package description:</h3><p>The package "huge" provides a general framework for high-dimensional undirected graph estimation. It integrates data preprocessing, neighborhood screening, graph estimation, and model selection techniques into a pipeline. In preprocessing stage, the nonparanormal(npn) transformation is applied to help relax the normality assumption. In the graph estimation stage, the graph structure is estimated by Meinshausen-Buhlmann graph estimation or the graphical lasso, and both methods can be further accelerated by the lossy screening rule preselecting the neighborhood of each variable by correlation thresholding. We target on high-dimensional data analysis usually d >> n, and the computation is memory-optimized using the sparse matrix output. We also provide a computationally efficient approach, correlation thresholding graph estimation. Three regularization/thresholding parameter selection methods are included in this package: (1)stability approach for regularization selection (2) rotation information criterion (3) extended Bayesian information criterion which is only available for the graphical lasso.</p>crantastic.orgtag:crantastic.org,2005:TimelineEvent/151312011-08-10T16:50:36Z2011-08-10T16:50:36Zhuge was upgraded to version 1.1.1<a href="/packages/huge">huge</a> was <span class="action">upgraded</span> to version <a href="/packages/huge/versions/13658">1.1.1</a><br /><h3>Package description:</h3><p>The package "huge" provides a general framework for high-dimensional undirected graph estimation. It integrates data preprocessing, neighborhood screening, graph estimation, and model selection techniques into a pipeline. In preprocessing stage, the nonparanormal(npn) transformation is applied to help relax the normality assumption. In the graph estimation stage, the graph structure is estimated by Meinshausen-Buhlmann graph estimation or the graphical lasso, and both methods can be further accelerated by the lossy screening rule preselecting the neighborhood of each variable by correlation thresholding. We target on high-dimensional data analysis usually d >> n, and the computation is memory-optimized using the sparse matrix output. We also provide a computationally efficient approach, correlation thresholding graph estimation. Three regularization/thresholding parameter selection methods are included in this package: (1)stability approach for regularization selection (2) rotation information criterion (3) extended Bayesian information criterion which is only available for the graphical lasso.</p>crantastic.orgtag:crantastic.org,2005:TimelineEvent/145882011-07-23T14:11:13Z2011-07-23T14:11:13Zhuge was upgraded to version 1.1.0<a href="/packages/huge">huge</a> was <span class="action">upgraded</span> to version <a href="/packages/huge/versions/13150">1.1.0</a><br /><h3>Package description:</h3><p>The package "huge" provides a general framework for high-dimensional undirected graph estimation. It integrates data preprocessing, neighborhood screening, graph estimation, and model selection techniques into a pipeline. In preprocessing stage, the nonparanormal(npn) transformation is applied to help relax the normality assumption. In the graph estimation stage, the graph structure is estimated by Meinshausen-Buhlmann graph estimation or the graphical lasso, and both methods can be further accelerated by the lossy screening rule preselecting the neighborhood of each variable by correlation thresholding. We target on high-dimensional data analysis usually d >> n, and the computation is memory-optimized using the sparse matrix output. We also provide a computationally efficient approach, correlation thresholding graph estimation. Three regularization/thresholding parameter selection methods are included in this package: (1)stability approach for regularization selection (2) rotation information criterion (3) extended Bayesian information criterion which is only available for the graphical lasso.</p>crantastic.orgtag:crantastic.org,2005:TimelineEvent/140962011-06-17T06:50:31Z2011-06-17T06:50:31Zhuge was upgraded to version 1.0.3<a href="/packages/huge">huge</a> was <span class="action">upgraded</span> to version <a href="/packages/huge/versions/12776">1.0.3</a><br /><h3>Package description:</h3><p>The package "huge" provides a general framework for high-dimensional undirected graph estimation. It integrates data preprocessing (Gaussianization), neighborhood screening, graph estimation, and model selection techniques into a pipeline. In preprocessing stage, the NonparaNormal(NPN) transformation is applied to help relax the normality assumption. In the graph estimation stage, the graph structure is estimated by the Meinshausen & Buhlmann Graph Estimation via Lasso (MBGEL) by default and it can be further accelerated by the Graph SURE Screening (GSS) subroutine which preselects the graph neighborhood of each variable. In the case d >> n, the computation is memory optimized and is targeted on larger-scale problems. We also provide two alternative approaches for the graph estimation stage:(1) Graph Estimation via Correlation Thresholding (GECT) which is highly efficient and (2) A slightly modified Graphical Lasso (GLASSO) procedure in which the memory usage is optimized using sparse matrix output. Three regularization/thresholding parameter selection methods are included in this package: (1) StARS: Stability Approach for Regularization Selection (2) RIC: Rotation Information Criterion (3) Extended Bayesian Information Criterion (EBIC only for GLASSO).</p>crantastic.orgtag:crantastic.org,2005:TimelineEvent/140512011-06-15T18:11:20Z2011-06-15T18:11:20Zhuge was upgraded to version 1.0.2<a href="/packages/huge">huge</a> was <span class="action">upgraded</span> to version <a href="/packages/huge/versions/12731">1.0.2</a><br /><h3>Package description:</h3><p>The package "huge" provides a general framework for high-dimensional undirected graph estimation. It integrates data preprocessing (Gaussianization), neighborhood screening, graph estimation, and model selection techniques into a pipeline. In preprocessing stage, the NonparaNormal(NPN) transformation is applied to help relax the normality assumption. In the graph estimation stage, the graph structure is estimated by the Meinshausen & Buhlmann Graph Estimation via Lasso (MBGEL) by default and it can be further accelerated by the Graph SURE Screening (GSS) subroutine which preselects the graph neighborhood of each variable. In the case d >> n, the computation is memory optimized and is targeted on larger-scale problems. We also provide two alternative approaches for the graph estimation stage:(1) Graph Estimation via Correlation Thresholding (GECT) which is highly efficient and (2) A slightly modified Graphical Lasso (GLASSO) procedure in which the memory usage is optimized using sparse matrix output. Three regularization/thresholding parameter selection methods are included in this package: (1) StARS: Stability Approach for Regularization Selection (2) RIC: Rotation Information Criterion (3) Extended Bayesian Information Criterion (EBIC only for GLASSO).</p>crantastic.orgtag:crantastic.org,2005:TimelineEvent/128672011-04-11T06:52:28Z2011-04-11T06:52:28Zhuge was upgraded to version 1.0.1<a href="/packages/huge">huge</a> was <span class="action">upgraded</span> to version <a href="/packages/huge/versions/11819">1.0.1</a><br /><h3>Package description:</h3><p>The package "huge" provides a general framework for high-dimensional undirected graph estimation. It integrates data preprocessing (Gaussianization), neighborhood screening, graph estimation, and model selection techniques into a pipeline. In preprocessing stage, the NonparaNormal(NPN) transformation is applied to help relax the normality assumption. In the graph estimation stage, the graph structure is estimated by the Meinshausen & Buhlmann Graph Estimation via Lasso (MBGEL) by default and it can be further accelerated by the Graph SURE Screening (GSS) subroutine which preselects the graph neighborhood of each variable. In the case d >> n, the computation is memory optimized and is targeted on larger-scale problems (d>6000). We also provide two alternative approaches for the graph estimation stage:(1) Graph Estimation via Correlation Thresholding (GECT) which is highly efficient and (2) A slightly modified Graphical Lasso (GLASSO) procedure in which the memory usage is optimized using sparse matrix output. Three regularization/thresholding parameter selection methods are included in this package: (1) StARS: Stability Approach for Regularization Selection (2) RIC: Rotation Information Criterion (3) Extended Bayesian Information Criterion (EBIC only for GLASSO).</p>crantastic.orgtag:crantastic.org,2005:TimelineEvent/120102011-03-02T17:50:21Z2011-03-02T17:50:21Zhuge was upgraded to version 1.0<a href="/packages/huge">huge</a> was <span class="action">upgraded</span> to version <a href="/packages/huge/versions/11098">1.0</a><br /><h3>Package description:</h3><p>The package "huge" provides a general framework for high-dimensional undirected graph estimation. It integrates data preprocessing (Gaussianization), neighborhood screening, graph estimation, and model selection techniques into a pipeline. In preprocessing stage, the NonparaNormal(NPN) transformation is applied to help relax the normality assumption. In the graph estimation stage, the graph structure is estimated by the Meinshausen & Buhlmann Graph Estimation via Lasso (MBGEL) by default and it can be further accelerated by the Graph SURE Screening (GSS) subroutine which preselects the graph neighborhood of each variable. In the case d >> n, the computation is memory optimized and is targeted on larger-scale problems (d>6000). We also provide two alternative approaches for the graph estimation stage:(1) Graph Estimation via Correlation Thresholding (GACT) which is highly efficient and (2) A slightly modified Graphical Lasso (GLASSO) procedure in which the memory usage is optimized using sparse matrix output. Three regularization/thresholding parameter selection methods are included in this package: (1) StARS: Stability Approach for Regularization Selection (2) PIC: Permutation Information Criterion (3) Extended Bayesian Information Criterion (EBIC only for GLASSO).</p>crantastic.orgtag:crantastic.org,2005:TimelineEvent/115042011-02-13T16:30:32Z2011-02-13T16:30:32Zhuge was upgraded to version 0.9.1<a href="/packages/huge">huge</a> was <span class="action">upgraded</span> to version <a href="/packages/huge/versions/10732">0.9.1</a><br /><h3>Package description:</h3><p>The package "huge" provides a general framework for high-dimensional undirected graph estimation. It integrates data preprocessing (Gaussianization), neighborhood screening, graph estimation, and model selection techniques into a pipeline. In preprocessing stage, the NonparaNormal(NPN) transformation is applied to help relax the normality assumption. In the graph estimation stage, the structure of either the whole graph or a pre-specified sub-graph is estimated by the Meinshausen & Buhlmann Graph Estimation via Lasso (GEL) by default and it can be further accelerated by the Graph SURE Screening (GSS) subroutine which preselects the graph neighborhood of each variable. In the case d >> n, the computation is memory optimized and is targeted on larger-scale problems (with d>2000). We also provide two alternative approaches for the graph estimation stage:(1) Graph Approximation via Correlation Thresholding (GACT) which is highly efficient and (2) A slightly modified Graphical Lasso (GLASSO) procedure in which the memory usage is optimized using sparse matrix output. Three regularization/thresholding parameter selection methods are included in this package: (1) StARS: Stability Approach for Regularization Selection (2) PIC: Permutation Information Criterion (3) Extended Bayesian Information Criterion (EBIC).</p>crantastic.orgtag:crantastic.org,2005:TimelineEvent/99512010-11-22T08:10:19Z2010-11-22T08:10:19Zhuge was upgraded to version 0.9<a href="/packages/huge">huge</a> was <span class="action">upgraded</span> to version <a href="/packages/huge/versions/9623">0.9</a><br /><h3>Package description:</h3><p>The package "huge" provides a general framework for high-dimensional undirected graph estimation. It integrates data preprocessing (Gaussianization), neighborhood screening, graph estimation, and model selection techniques into a pipeline. In preprocessing stage, the NonparaNormal(NPN) transformation is applied to help relax the normality assumption. In the graph estimation stage, the structure of either the whole graph or a pre-specified sub-graph is estimated by the Meinshausen & Buhlmann Graph Estimation via Lasso (GEL) by default and it can be further accelerated by the Graph SURE Screening (GSS) subroutine which preselects the graph neighborhood of each variable. In the case d >> n, the computation is memory optimized and is targeted on larger-scale problems (with d>2000). We also provide two alternative approaches for the graph estimation stage:(1) Graph Approximation via Correlation Thresholding (GACT) which is highly efficient and (2) A slightly modified Graphical Lasso (GLASSO) procedure in which the memory usage is optimized using sparse matrix output. Three regularization/thresholding parameter selection methods are included in this package: (1) StARS: Stability Approach for Regularization Selection (2) PIC: Permutation Information Criterion (3) Extended Bayesian Information Criterion (EBIC).</p>crantastic.org