observe_ridge

function observe_ridge(W:Real[_,_], σ2:Real[_], N:Real[_,_], Λ:LLT, α:Real, γ:Real[_]) -> Real

Observe ridge regression parameters.

  • W: Weight matrix.
  • σ2: Variance vector.
  • N: Prior precision times mean for weights, where each column represents the mean of the weight for a separate output.
  • Λ: Common prior precision.
  • α: Common prior weight and likelihood covariance shape.
  • β: Prior covariance scale accumulators.

Returns: Matrix of weights and vector of variances, where each column in the matrix and element in the vector corresponds to a different output in the regression.