log-loss
An evaluation metric commonly used for binary classification prediction models, defined as the negative log-likelihood of the predicted probabilities under the true labels; it penalizes both misclassification and overconfident probability estimates and is sensitive to the quality of model calibration.
Timeline (last 90 days)
Physiological trajectory features (R2 relative to R1) brought consistent improvement in external prediction on eICU, with a log-loss difference of -0.00922 (95% CI -0.01148 to -0.00704).
When the number of observations was matched, spreading measurements across the entire time window compared with random retention gave a log-loss improvement of 0.03280 for R3, versus only 0.00703 for R2.
Log-loss is a commonly used evaluation metric for binary classification prediction models, defined as the negative log-likelihood of the predicted probabilities under the true labels; lower values are better, and it is sensitive to the quality of model calibration.
All Facts (3)
Physiological trajectory features (R2 relative to R1) brought consistent improvement in external prediction on eICU, with a log-loss difference of -0.00922 (95% CI -0.01148 to -0.00704).
50%UnverifiedWhen the number of observations was matched, spreading measurements across the entire time window compared with random retention gave a log-loss improvement of 0.03280 for R3, versus only 0.00703 for R2.
50%UnverifiedLog-loss is a commonly used evaluation metric for binary classification prediction models, defined as the negative log-likelihood of the predicted probabilities under the true labels; lower values are better, and it is sensitive to the quality of model calibration.
50%