Variational Garrote with automatic differentiation yields more robust sparse variable selection than Ridge and LASSO in very sparse regimes, and the sharp rise in selection uncertainty marks the true number of relevant variables.
To address this, we introduce a unified measure for variable sparsity that can be applied consistently across all three models of Ridge, LASSO, and VG
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Variational Garrote for Statistical Physics-based Sparse and Robust Variable Selection
Variational Garrote with automatic differentiation yields more robust sparse variable selection than Ridge and LASSO in very sparse regimes, and the sharp rise in selection uncertainty marks the true number of relevant variables.