Sparse Generalized Eigenvalue Problem Via Smooth Optimization

Junxiao Song, Prabhu Babu, Daniel P. Palomar
2015 IEEE Transactions on Signal Processing  
In this paper, we consider an ℓ_0-norm penalized formulation of the generalized eigenvalue problem (GEP), aimed at extracting the leading sparse generalized eigenvector of a matrix pair. The formulation involves maximization of a discontinuous nonconcave objective function over a nonconvex constraint set, and is therefore computationally intractable. To tackle the problem, we first approximate the ℓ_0-norm by a continuous surrogate function. Then an algorithm is developed via iteratively
more » ... ing the surrogate function by a quadratic separable function, which at each iteration reduces to a regular generalized eigenvalue problem. A preconditioned steepest ascent algorithm for finding the leading generalized eigenvector is provided. A systematic way based on smoothing is proposed to deal with the "singularity issue" that arises when a quadratic function is used to majorize the nondifferentiable surrogate function. For sparse GEPs with special structure, algorithms that admit a closed-form solution at every iteration are derived. Numerical experiments show that the proposed algorithms match or outperform existing algorithms in terms of computational complexity and support recovery.
doi:10.1109/tsp.2015.2394443 fatcat:35x6gxlvwjhjhg6s26jninbrje