Algorithms for Sparse Support Vector Machines
Many problems in classification involve huge numbers of irrelevant features. Model selection reveals the crucial features, reduces the dimensionality of feature space, and improves model interpretation. In the support vector machine literature, model selection is achieved by ℓ_1 penalties. These convex relaxations seriously bias parameter estimates toward 0 and tend to admit too many irrelevant features. The current paper presents an alternative that replaces penalties by sparse-set constraints. Penalties still appear, but serve a different purpose. The proximal distance principle takes a loss function L(β) and adds the penalty ρ/2dist(β, S_k)^2 capturing the squared Euclidean distance of the parameter vector β to the sparsity set S_k where at most k components of β are nonzero. If β_ρ represents the minimum of the objective f_ρ(β)=L(β)+ρ/2dist(β, S_k)^2, then β_ρ tends to the constrained minimum of L(β) over S_k as ρ tends to ∞. We derive two closely related algorithms to carry out this strategy. Our simulated and real examples vividly demonstrate how the algorithms achieve much better sparsity without loss of classification power.
READ FULL TEXT