initial.convex: Convex Relaxation for Sparse GEP

Description Usage Arguments Value Author(s) References

Description

Estimate the K-dimensional subspace spanned by the largest K generalized eigenvector by solving a convex relaxation. The details are given in Tan et al. (2018).

Usage

1
initial.convex(A, B, lambda, K, nu = 1, epsilon = 0.005, maxiter = 1000, trace = FALSE)

Arguments

A

Input the matrix A for sparse generalized eigenvalue problem.

B

Input the matrix B for sparse generalized eigenvalue problem.

lambda

A positive tuning parameter that constraints the solution to be sparse

K

A positive integer tuning parameter that constraints the solution to be low rank.

nu

An ADMM tuning parameter that controls the convergence of the ADMM algorithm.

epsilon

Threshold for convergence. Default value is 0.005.

maxiter

Maximum number of iterations. Default is 1000 iterations.

trace

Default value of trace=FALSE. If trace=TRUE, each iteration of the ADMM algorithm is printed.

Value

Pi

Estimated subspace Pi

Author(s)

Kean Ming Tan

References

Sparse Generalized Eigenvalue Problewm: Optimal Statistical Rates via Truncated Rayleigh Flow", by Tan et al. (2018). To appear in Journal of the Royal Statistical Society: Series B. https://arxiv.org/pdf/1604.08697.pdf.


rifle documentation built on May 2, 2019, 2:51 p.m.