Metrics
Metrics#
- class torchmimic.metrics.AUCROC(average=None)#
AUCROC scoring class
- class torchmimic.metrics.AverageMeter#
Class used to collect values and return a running average
- reset()#
Resets private members
- update(val, _n=1)#
Updates class members
- Parameters
val (float) – value used to update running average
_n (int) – sample size used to calculate value
- class torchmimic.metrics.MetricMeter(score_fn)#
Class used to collect values and evaluate them using a scoring function
- reset()#
Resets private members
- score()#
Scores true and predicted values :returns: the output of the score function given the predicted and true labels :rtype: int
- update(true, pred)#
Updates list of true and predicted values
- Parameters
true (np.array) – true labels
pred (np.array) – predicted labels
- torchmimic.metrics.accuracy(true, pred)#
Returns the accuracy for the provided true and predicted values
- Parameters
true (np.array) – true values
pred (np.array) – predicted values
- Returns
accuracy score
- Return type
int
- torchmimic.metrics.aucpr(true, pred)#
Returns the AUC-PR for the provided true and predicted values
- Parameters
true (np.array) – true values
pred (np.array) – predicted values
- Returns
AUC-PR score
- Return type
int
- torchmimic.metrics.balanced_accuracy(true, pred)#
Returns the Balanced Accuracy for the provided true and predicted values
- Parameters
true (np.array) – true values
pred (np.array) – predicted values
- Returns
Balanced Accuracy score
- Return type
int
- torchmimic.metrics.f1(true, pred)#
Returns the F1-score for the provided true and predicted values
- Parameters
true (np.array) – true values
pred (np.array) – predicted values
- Returns
F1-score
- Return type
int
- torchmimic.metrics.kappa(true, pred)#
Returns the Cohen’s Kappa for the provided true and predicted values
- Parameters
true (np.array) – true values
pred (np.array) – predicted values
- Returns
Cohen’s Kappa score
- Return type
int
- torchmimic.metrics.mae(true, pred)#
Returns the Mean Absolute Error/Deviation for the provided true and predicted values
- Parameters
true (np.array) – true values
pred (np.array) – predicted values
- Returns
MAE/MAD score
- Return type
int