boax.acquisitions.log_probability_of_improvement#
- boax.acquisitions.log_probability_of_improvement(bounds, improvement_factor=1.0, num_raw_samples=512, num_restarts=10)#
The Log Probability of Improvement acquisition function.
Logarithm of the probability of improvement over the best function value observed so far.
logPI(x) = log(P(y >= best)), y ~ f(x)
Example
>>> acqf = log_probability_of_improvement(0.2, model) >>> log_poi = acqf(index_points)
- Parameters:
best – The best function value observed so far.
model – The surrogate model.
- Return type:
- Returns:
The corresponding Acquisition.