Elastic net
Also: Elastic Net, elastic-net regularization
A regularizer that adds both λ₁·Σ|wᵢ| and λ₂·Σwᵢ² to the loss (often written with a mixing ratio α). The L1 part zeroes out irrelevant features; the L2 part makes groups of correlated features be kept or dropped together.
L_total = L_data + λ₁ Σ |wᵢ| + λ₂ Σ wᵢ²
Example
With two nearly identical gene features, lasso tends to keep one at random and zero the other, while elastic net with a 50/50 mix keeps both with similar, shrunken weights.