09e20914e3b49c264cb41262ea24fd54121e0b84,plot_poldrack_space_net.py,,,#,6

Before Change

////// Fit and predict ////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
from nilearn.decoding import SpaceNetRegressor
penalties = ["Smooth-LASSO", "TV-L1"]
decoders = {}
for penalty in penalties:
    decoder = SpaceNetRegressor(memory=mem, mask=mask_img, verbose=2,
                                penalty=penalty)
    decoder.fit(X_train, y_train)  // fit
    decoders[penalty] = decoder

////// Visualization //////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
import matplotlib.pyplot as plt
from nilearn.image import mean_img
from nilearn.plotting import plot_stat_map
background_img = mean_img(X_train)

After Change

////// Fit and predict ////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
import os
from nilearn.decoding import SpaceNetRegressor
penalty = "TV-L1"
l1_ratio = .3
alpha = None
decoder = SpaceNetRegressor(memory=mem, mask=mask_img, verbose=2,
                            n_jobs=int(os.environ.get("N_JOBS", 1)),
                            l1_ratio=l1_ratio, penalty=penalty, alpha=alpha,
                            screening_percentile=100., tol=1e-8,
                            max_iter=1000)
decoder.fit(X_train, y_train)  // fit
coef_niimg = decoder.coef_img_
coef_niimg.to_filename("poldrack_%s(l1_ratio=%g, alpha=%s)_weights.nii" % (
        penalty, l1_ratio, alpha))

////// Visualization //////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////

Instances