Deep Learning with Keras

An overview of losses functions

Losses functions (or objective functions, or optimization score function; for more information, refer to https://keras.io/losses/) can be classified into four categories:

• Accuracy which is used for classification problems. There are multiple choices: binary_accuracy (mean accuracy rate across all predictions for binary classification problems), categorical_accuracy (mean accuracy rate across all predictions for multiclass classification problems), sparse_categorical_accuracy (useful for sparse targets), and top_k_categorical_accuracy (success when the target class is within the top_k predictions provided).
• Error loss, which measures the difference between the values predicted and the values actually observed. ...