INNER CODE UNIT · Python

n_jobs

epigen/MrBiomics · workflow/scripts/crossprediction.py:65

    n_jobs=-1,
    random_state=42,
    verbose=1,
    l1_ratio=0.5
)

#################################################################
#### Get connectivity matrix from LOO-CV strategy prediction probabilities
#################################################################

classnames, groups = np.unique(y, return_inverse=True)
cv = sklearn.model_selection.LeaveOneGroupOut()
pred = sklearn.model_selection.cross_val_predict(estimator=clf, X=X, y=y, groups=groups, cv=cv, n_jobs=-1, method='predict_proba')

# sanity check if approach works as intended -> pred. prob for correct label has to be 0!
result=pd.DataFrame(pred, columns=classnames, index=y)
for index, row in result.iterrows():
    if row[index]!=0:

View source record →