Bayesian optimization is a strategy to find the global optimum in the hyperparameter space of your PHOTONAI-Pipeline. Beside SMAC3, Scikit-Optimize is one of the supported bayesian optimization frameworks. For a detailed documentation have a look at Scikit-Optimize.

To enable Scikit-Optimize in your Hyperpipe just add it as shown here:

my_pipe = Hyperpipe('Scikit-optimizer_pipeline',
                    optimizer_params={'n_configurations': 25,
                                      'acq_func': 'LCB',
                                      'acq_func_kwargs': {'kappa': 1.96}

Set: optimizer = 'sk_opt'

Parameter Type Description
n_configurations int Number of configurations to test in every outer_fold. Default is set to 25.
acq_func str Function for minimizing over the gaussian prior. Can be either ['LCB', 'EI', 'PI', 'gp_hedge', 'EIps', 'PIps']. Default is 'gp_hedge'.
acq_func_kwargs dict Kwargs depending on acq_func. Possible settings can be found under: Scikit-Optimize.