L1Regularization<T, TInput, TOutput>

Models & Types ยท AiDotNet.Regularization

Implements L1 regularization (also known as Lasso), a technique that adds a penalty equal to the absolute value of the magnitude of coefficients.

Inheritance: RegularizationBase<T, TInput, TOutput> โ†’ L1Regularization<T, TInput, TOutput>

For Beginners

L1 regularization helps create simpler models by completely removing less important features.

Think of it like a strict budget committee:

  • It forces the model to focus only on the most important features
  • Less important features get their coefficients reduced to exactly zero
  • This means some features are completely eliminated from the model

How It Works

L1 regularization adds a penalty term to the loss function equal to the sum of the absolute values of the model coefficients, multiplied by a regularization strength parameter. This encourages sparse models by driving some coefficients to exactly zero, effectively performing feature selection.

Constructors

Constructor Summary
L1Regularization(RegularizationOptions options) Initializes a new instance of the L1Regularization class with optional custom options.

Methods

Method Summary
Matrix<T> Regularize(Matrix<T> data) Applies L1 regularization to a vector.
TOutput Regularize(TOutput gradient, TOutput coefficients)
Vector<T> Regularize(Vector<T> data) Applies L1 regularization to a vector.
Vector<T> Regularize(Vector<T> gradient, Vector<T> coefficients)