ridge_regression

torchhd.ridge_regression(samples: Tensor, labels: Tensor, alpha: float | None = 1)[source]

Compute weights (readout matrix) with ridge regression.

It is a common way to form classifiers within randomized neural networks see, e.g., Randomness in Neural Networks: An Overview.

Parameters:

  • samples (Tensor) – The feature vectors.

  • labels (Tensor) – The target vectors, typically one-hot vectors for classification problems.

  • alpha (float, optional) – Scalar for the variance of the samples. Default is 1.

Shapes:

  • Samples: \((n, d)\)

  • Labels: \((n, c)\)

  • Output: \((c, d)\)