CrossEntropy¶
Cross entropy loss.
This is a generalization of logistic loss to multiple classes.
Parameters¶
-
class_weight (Dict[Union[bool, str, int], float]) – defaults to
NoneA dictionary that indicates what weight to associate with each class.
Examples¶
>>> from river import optim
>>> y_true = [0, 1, 2, 2]
>>> y_pred = [
... {0: 0.29450637, 1: 0.34216758, 2: 0.36332605},
... {0: 0.21290077, 1: 0.32728332, 2: 0.45981591},
... {0: 0.42860913, 1: 0.33380113, 2: 0.23758974},
... {0: 0.44941979, 1: 0.32962558, 2: 0.22095463}
... ]
>>> loss = optim.losses.CrossEntropy()
>>> for yt, yp in zip(y_true, y_pred):
... print(loss(yt, yp))
1.222454
1.116929
1.437209
1.509797
>>> for yt, yp in zip(y_true, y_pred):
... print(loss.gradient(yt, yp))
{0: -0.70549363, 1: 0.34216758, 2: 0.36332605}
{0: 0.21290077, 1: -0.67271668, 2: 0.45981591}
{0: 0.42860913, 1: 0.33380113, 2: -0.76241026}
{0: 0.44941979, 1: 0.32962558, 2: -0.77904537}
Methods¶
call
Returns the loss.
Parameters
- y_true
- y_pred
Returns
The loss(es).
clone
Return a fresh estimator with the same parameters.
The clone has the same parameters but has not been updated with any data. This works by looking at the parameters from the class signature. Each parameter is either - recursively cloned if it's a River classes. - deep-copied via copy.deepcopy if not. If the calling object is stochastic (i.e. it accepts a seed parameter) and has not been seeded, then the clone will not be idempotent. Indeed, this method's purpose if simply to return a new instance with the same input parameters.
gradient
Return the gradient with respect to y_pred.
Parameters
- y_true
- y_pred
Returns
The gradient(s).
mean_func
Mean function.
This is the inverse of the link function. Typically, a loss function takes as input the raw output of a model. In the case of classification, the raw output would be logits. The mean function can be used to convert the raw output into a value that makes sense to the user, such as a probability.
Parameters
- y_pred
Returns
The adjusted prediction(s).