http://app.cul.columbia.edu:8080/ac/handle/10022/AC:P:29741 Search the Channel search http://app.cul.columbia.edu:8080/ac/simple-search http://app.cul.columbia.edu:8080/ac/handle/10022/AC:P:29744 Title: Priors on the Variance in Sparse Bayesian Learning; the demi-Bayesian Lasso <br/> <br/>Author(s): Madigan, David <br/> <br/>Abstract: We explore the use of proper priors for variance parameters of certain sparse Bayesian regression models. This leads to a connection between sparse Bayesian learning (SBL) models (Tipping, 2001) and the recently proposed Bayesian Lasso (Park and Casella, 2008). We outline simple modifications of existing algorithms to solve this new variant which essentially uses type-II maximum likelihood to fit the Bayesian Lasso model. We also propose an Elastic-net (Zou and Hastie, 2005) heuristic to help with modeling correlated inputs. Experimental results show the proposals to compare favorably to both the Lasso and traditional and more recent sparse Bayesian algorithms. Mon, 29 Oct 2007 22:58:59 GMT http://app.cul.columbia.edu:8080/ac/handle/10022/AC:P:29743 Title: Tree-Based Integration of One-versus-Some (OVS) Classifiers for Multiclass Classification <br/> <br/>Author(s): Ding, Yuejing; Zheng, Tian <br/> <br/>Abstract: Motivated by applications such as gene expression analysis, binary classification has achieved notable success. (e.g., cancer samples versus normal samples) When comes to multiclass classification, the extension is not straightforward. There has been two main directions on such extensions: 1) via a sequence of nested binary classifiers in a classification tree or 2) via classifier ensembles that integrate votes from all one-versus-all (OVA) classifiers or all all-pairs (AP) classifiers. In this article, we present a new way to combine both strategies in a multiclass classification. <br/> <br/>Description: This technical report was included in the Joint Statistical Meeting 2006 proceedings. Sat, 29 Oct 2005 22:58:59 GMT