boax.acquisitions.expected_improvement

boax.acquisitions.expected_improvement#

boax.acquisitions.expected_improvement(bounds, improvement_factor=1.0, num_raw_samples=512, num_restarts=10)#

The Expected Improvement acquisition function.

Expected improvement over the best function value observed so far.

EI(x) = E(max(f(x) - best, 0)),

where the expectation is taken over the value of stochastic function f at x.

Example

>>> acqf = expected_improvement(0.2, model)
>>> ei = acqf(index_points)
Parameters:
  • best – The best function value observed so far.

  • model – The surrogate model.

Return type:

Acquisition

Returns:

The corresponding Acquisition.