Abstract:This paper proposes the nonnegative adaptive lasso method for variable selection both in the classical fixed p setting (OLS initial estimator) and the ultra-high dimensional setting (root-n-consistent initial estimator). This method is an extension of the adaptive lasso with nonnegative constraint on the coefficients. It is shown to have asymptotic unbiasedness, asymptotic normality and variable selection consistency and its mean squared error decays fast too. Comparing with other procedures, nonnegative adaptive lasso satisfies oracle properties and can select the true variables under fewer assumptions. To get the solution of the nonnegative adaptive lasso, we extend the multiplicative approach for computing. This algorithm is valid for the general framework where the number of regression parameters p is allowed to very large. Simulations are performed to illustrate above results.The constrained index tracking problem in the stock market without short sales is studied in the empirical part. A two-stage method, nonnegative adaptive lasso+nonnegative LS, is applied in the financial modeling. The tracking results indicate that nonnegative adaptive lasso and the two-stage method can both get small tracking error and is successful in assets selection. (C) 2016 Elsevier B.V. All rights reserved.

Asymptotic properties of adaptive Dantzig selector

Adaptive Lasso in high-dimensional settings

Parallelism, uniqueness, and large-sample asymptotics for the Dantzig selector

The constrained Dantzig selector with enhanced consistency

Adaptive $$l_p$$ $$(0<p<1)$$ Regularization: Oracle Property and Applications

Asymptotic Behavior for S-estimators in Random Design Linear Model with Long-Range-dependent Errors

Simultaneous analysis of Lasso and Dantzig selector

High-dimensional stochastic optimization with the generalized Dantzig estimator

Precise asymptotics of partial sums for multidimensionally indexed random variables.

Two step estimations via the Dantzig selector for models of stochastic processes with high-dimensional parameters

On asymptotics of eigenvectors of large sample covariance matrix

The Dantzig selector for a linear model of diffusion processes

ASYMPTOTIC PROPERTIES OF ADAPTIVE DESIGNS VIA STRONG APPROXIMATIONS

Discussion: The Dantzig selector: Statistical estimation when $p$ is much larger than $n$

Asymptotic Theory of Least Distances Estimate in Multivariate Linear Models

Asymptotics for A Class of Dependent Random Variables

Post Selection Shrinkage Estimation for High Dimensional Data Analysis

The Dantzig selector for diffusion processes with covariates

The Adaptive Lasso and Its Oracle Properties

A Note on Adaptive Lp Regularization

Nonnegative Adaptive Lasso for Ultra-High Dimensional Regression Models and A Two-Stage Method Applied in Financial Modeling