Efficient regularized regression with L₀ penalty for variable selection and network construction
Summary: Variable selections for regression with high-dimensional big data have found many applications in bioinformatics and computational biology. One appealing approach is the \(L_0\) regularized regression which penalizes the number of nonzero features in the model directly. However, it is well known that \(L_0\) optimization is NP-hard and computationally challenging. In this paper, we propose efficient EM (\(L_0\)EM) and dual \(L_0\)EM (D\(L_0\)EM) algorithms that directly approximate the \(L_0\) optimization problem. While \(L_0\)EM is efficient with large sample size, D\(L_0\)EM is efficient with high-dimensional (\(n \ll m\)) data. They also provide a natural solution to all \(L_p\)\ \ \(p \in [0,2]\) problems, including lasso with \(p = 1\) and elastic net with \(p \in [1,2]\). The regularized parameter \(\lambda\) can be determined through cross validation or AIC and BIC. We demonstrate our methods through simulation and high-dimensional genomic data. The results indicate that \(L_0\) has better performance than lasso, SCAD, and MC+, and \(L_0\) with AIC or BIC has similar performance as computationally intensive cross validation. The proposed algorithms are efficient in identifying the nonzero variables with less bias and constructing biologically important networks with high-dimensional big data.
- Variable selection and estimation using a continuous approximation to the \(L_0\) penalty
- scientific article; zbMATH DE number 6162361
- scientific article; zbMATH DE number 6982301
- Variable selection and estimation in generalized linear models with the seamless L₀ penalty
- Regularization and Variable Selection Via the Elastic Net
- A generic path algorithm for regularized statistical estimation
- A new look at the statistical model identification
- Estimating the dimension of a model
- scientific article; zbMATH DE number 845714 (Why is no real title available?)
- scientific article; zbMATH DE number 6162361 (Why is no real title available?)
- Improved iteratively reweighted least squares for unconstrained smoothed _q minimization
- Iteratively reweighted least squares minimization for sparse recovery
- Nearly unbiased variable selection under minimax concave penalty
- On the adaptive elastic net with a diverging number of parameters
- Partial correlation estimation by joint sparse regression models
- Sparse Approximation via Penalty Decomposition Methods
- SparseNet: coordinate descent with nonconvex penalties
- Stability selection. With discussion and authors' reply
- The Adaptive Lasso and Its Oracle Properties
- The risk inflation criterion for multiple regression
- Tuning parameter selection in high dimensional penalized likelihood
- Variable Selection via Nonconcave Penalized Likelihood and its Oracle Properties
- Broken adaptive ridge regression and its asymptotic properties
- Simultaneous estimation and variable selection for incomplete event history studies
- Network-based penalized regression with application to genomic data
- Simultaneous Estimation and Variable Selection for Interval-Censored Data With Broken Adaptive Ridge Regression
- Incorporating Predictor Network in Penalized Regression with Application to Microarray Data
- Likelihood-based selection and sharp parameter estimation
- L0-Regularized Learning for High-Dimensional Additive Hazards Regression
- Fast best subset selection: coordinate descent and local combinatorial optimization algorithms
- Truncated \(L_1\) regularized linear regression: theory and algorithm
- Simultaneous variable selection and estimation for joint models of longitudinal and failure time data with interval censoring
- Variables selection using \(\mathcal{L}_0\) penalty
- A network Lasso model for regression
- Estimation of l₀ norm penalized models: a statistical treatment
- Variable selection for misclassified current status data under the proportional hazards model
- Algorithmic geolocation of harvest in hand-picked agriculture
- A conditional approach for regression analysis of case \(K\) interval-censored failure time data with informative censoring
- Variable selection for high-dimensional partly linear additive Cox model with application to Alzheimer's disease
- Variable selection for bivariate interval-censored failure time data under linear transformation models
- Penalized variable selection with broken adaptive ridge regression for semi-competing risks data
- Adaptive ridge approach to heteroscedastic regression
- Simultaneous Coefficient Clustering and Sparsity for Multivariate Mixed Models
- Simultaneous variable selection and estimation of mixed panel count data with informative observation processes
- Variable selection in the joint frailty model of recurrent and terminal events using broken adaptive ridge regression
- Variable selection with broken adaptive ridge regression for interval-censored competing risks data
- Variable selection in causal semiparametric transformation models with all-or-nothing treatment compliance
- An attention algorithm for solving large scale structured \(l_0\)-norm penalty estimation problems
- A scalable surrogate L₀ sparse regression method for generalized linear models with applications to large scale data
This page was built for publication: Efficient regularized regression with \(L_0\) penalty for variable selection and network construction
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2011726)
