def compile_pain_decoder(self, pain_dir=None):
print("compiling pain dir")
if os.path.isfile(self.decoder_file):
print("pain data pre-saved, loading...")
stats = pickle.load(open(self.decoder_file, "rb"))
# pdata = Brain_Data(self.decoder_file)
else:
print("pain data doesn't exist, getting...")
if (pain_dir is None):
pdata = fetch_pain()
else:
pdata = fetch_pain(pain_dir)
pdata.Y = pdata.X["PainLevel"]
stats = pdata.predict(algorithm='ridge', plot=False)
with open(self.decoder_file, "wb") as f:
pickle.dump(stats, f, pickle.HIGHEST_PROTOCOL)
self.decoder_origin = 'chang_pain_data'
print("pain data loaded.")
#train_subjs=pandas.DataFrame(data_subjs).sample(frac=0.9,random_state=2057)
#test_subjs= [s for s in data_subjs if s not in train_subjs[0]]
#data_train = pdata[train_subjs[0]]
#data_test = pdata[test_subjs]
#stats = data_train.predict(algorithm='ridge',plot=False)
self.stats = stats
self.decoder = stats['weight_map']
cross-validation parameters. Currently, we have several different linear algorithms
implemented from `scikit-learn <http://scikit-learn.org/stable/>`_.
"""
#########################################################################
# Load Data
# ---------
#
# First, let's load the pain data for this example. We need to specify the
# training levels. We will grab the pain intensity variable from the data.X
# field.
from nltools.datasets import fetch_pain
data = fetch_pain()
data.Y = data.X['PainLevel']
#########################################################################
# Prediction with Cross-Validation
# --------------------------------
#
# We can now predict the output variable is a dictionary of the most
# useful output from the prediction analyses. The predict function runs
# the prediction multiple times. One of the iterations uses all of the
# data to calculate the 'weight_map'. The other iterations are to estimate
# the cross-validated predictive accuracy.
stats = data.predict(algorithm='ridge',
cv_dict={
'type': 'kfolds',
