def nonLinearInterpAupr(y_score, y_true):
if y_score.ndim > 1:
sub_stats = pd.DataFrame(np.concatenate((y_score, y_true), axis=1),
dtype='float64')
else:
sub_stats = pd.DataFrame(np.array([y_score, y_true]).T)
results = getAUROC_PR(sub_stats)
return results, y_score, y_true
def nonLinearInterpAupr(y_score, y_true):
""" Given ground truth targets and predicted targets, calculates
weighted non-linear interpolated AUPRC.
Parameters
----------
y_true (list-like) - actual target values
y_score (list-like) - predicted target values
average (str, default 'micro') - one of 'micro', 'macro', 'samples', 'weighted'
(see http://scikit-learn.org/stable/modules/generated/sklearn.metrics.roc_auc_score.html)
sample_weight (list-like, optional) - of shape (n_samples,)
Returns
-------
results (list), y_score (np.ndarray), y_true (np.ndarray)
"""
if y_score.ndim > 1:
sub_stats = pd.DataFrame(np.concatenate((y_score, y_true), axis=1),
dtype='float64')
else:
sub_stats = pd.DataFrame(np.array([y_score, y_true]).T)
results = getAUROC_PR(sub_stats)
return results, y_score, y_true