WeightedProportion.loss#

WeightedProportion.loss(W, X, y)[source]#

Compute root mean squared loss with data X, weights W, and predictor y.

Parameters:
  • W (ndarray) – Convex combination weights.

  • X (ndarray) – Donor matrix.

  • y (ndarray) – Treated unit outcomes in the pre-treatment period.

Return type:

float