Hands-On Deep Learning with R

Loss functions

With the preceding optimizers, you can select whichever one you like and try different algorithms to see which produces the best results. With the loss function, there are some choices that are more appropriate than others based on the problem being solved:

• binary_crossentropy: This loss function is used for classification problems where the requirement is to assign input data to one of two classes. This can also be used when assigning input data to more than one or two classes if it is possible for a given case to belong to more than one class. In this case, each target class is treated as a separate binary class (for each target class the given case belongs to or does not belong to). This can also be used for regression ...