def test_gp_std(get_kernel):
from uncoverml.optimise.models import TransformedGPRegressor
np.random.seed(10)
sklearn_gp = TransformedGPRegressor(kernel=get_kernel(length_scale=1))
sklearn_gp.fit(X=1 + np.random.rand(10, 3), y=1 + np.random.rand(10))
p, v, uq, lq = sklearn_gp.predict_dist(X=1 + np.random.rand(5, 3))
import uncoverml as ls
import uncoverml.config
import uncoverml.mllog
from uncoverml.config import ConfigException
from uncoverml.optimise.models import (TransformedGPRegressor, kernels,
TransformedSVR)
from uncoverml.scripts.uncoverml import _load_data
from uncoverml.transforms import target as transforms
from uncoverml.optimise.models import transformed_modelmaps as all_modelmaps
log = logging.getLogger(__name__)
pca = decomposition.PCA()
algos = {k: v(ml_score=True) for k, v in all_modelmaps.items()}
algos['transformedgp'] = TransformedGPRegressor(n_restarts_optimizer=10,
normalize_y=True,
ml_score=True)
algos['transformedsvr'] = TransformedSVR(verbose=True,
max_iter=1000000,
ml_score=True)
def setup_pipeline(config):
if config.optimisation['algorithm'] not in algos:
raise ConfigException('optimisation algo must exist in algos dict')
steps = []
param_dict = {}
if 'featuretransforms' in config.optimisation:
config.featuretransform = config.optimisation['featuretransforms']
if 'pca' in config.featuretransform:
steps.append(('pca', pca))