Component selection and smoothing in smoothing spline analysis of variance models -- COSSO

By Yi Lin, Hao Helen Zhang. 2003

140 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Fused Lasso Additive Model (2014, methodology): “It selects a subset of features to include in the model, and for these features it ﬁts piecewise constant functions with knots that are chosen adaptively based on the data.”
Using recursive algorithms for the efficient identification of smoothing spline ANOVA models (2010, methodology): “The COSSO (Lin and Zhang 2006) penalizes the sum of norms, using a Least Absolute Shrinkage and Selection Operator (LASSO)-type penalty (Tibshirani 1996) for the ANOVA model, which allows us to identify the informative predictor terms F j with an estimate of f that minimizes 1 N N k=1 z k − f (X k…”
Semiparametric regression models with additive nonparametric components and high dimensional parametric components (2012, methodology): “For example, Lin and Zhang (2006) have proposed the COmponent Selection and Smoothing Operator (COSSO) which can be viewed as a functional generalization of LASSO using the Sobolev norm as the penalty.”
Variable selection for support vector machines via smoothing spline anova (2006, methodology): “The penalty ∑q α=1 ‖P αf‖ is a sum of RKHS component norms, different from the squared RKHS norm penalty used in the standard SVM. Lin and Zhang (2002) proposed and studied this type of regularization in the penalized likelihood regression setting.”
Shrinkage Estimation of the Varying Coefficient Model (2009, background): “, varying coefficient models) is not trivial (Zhang and Lin 2003; Fan and Li 2004).”
Parsimonious additive models (2007, methodology): “In this approach, only one penalization term is needed to penalize linear and nonlinear components (Lin and Zhang, 2002).”
Moving beyond regression techniques in cardiovascular risk prediction: applying machine learning to address analytic challenges (2016, methodology): “While GAMs are not very efficient with many predictors, they have been combined with regularized methods such as LASSO to perform predictor selection.(29)”
Implementation and evaluation of nonparametric regression procedures for sensitivity analysis of computationally demanding models (2009, methodology): “It can be demonstrated analytically that the COSSO penalty with one tuning parameter gives as much flexibility as the smoothing spline penalty with p tuning parameters [40].”
ADAPTIVE SEMI-VARYING COEFFICIENT MODEL SELECTION (2012, background): “See for example Zhang and Lin (2003), Wang, Li, and Huang (2008), and Zou and Li (2008).”