Pérdida Logarítmica, also known as logistic loss or pérdida de entropía cruzada, is a performance metric used primarily in clasificación binaria problems where the predicted output is a probability value ranging between 0 and 1. It quantifies the difference between the predicted probabilities and the actual class labels (0 or 1). The objective of using log loss is to evaluate how well a classification model predicts probabilities for binary outcomes, with lower values indicating better rendimiento del modelo.
Matemáticamente, la pérdida logarítmica se calcula usando la siguiente fórmula:
Log Loss = -1/N * Σ [y * log(p) + (1 – y) * log(1 – p)]
Donde:
- N es el número total de predicciones.
- y es la etiqueta real (0 o 1).
- p es la probabilidad predicha de la clase positiva (1).
The Log Loss value ranges from 0 to infinity, where 0 indicates perfect predictions (model outputs probabilities of either 0 or 1 correctly) and larger values indicate worse performance. A model that predicts probabilities close to the true labels will have a lower log loss, while a model that predicts probabilities far las predicciones que se alejan de las etiquetas verdaderas incurrirán en una pérdida logarítmica mayor.
Log Loss is particularly useful in scenarios where the output is not just a hard classification but a probability, making it suitable for applications such as regresión logística, neural networks, and other probabilistic classifiers. It is also widely utilizado en aprendizaje automático competencias, como las organizadas por Kaggle, para evaluar el rendimiento del modelo.