Confidence Interval

To toggle the Prediction's Confidence Interval bounds of the in the chart, click in the Prediction's legend or line

Interval Bounds

When the Confidence Interval is active, the Prediction will display the 50% and 90% bounds (upper and lower)

MAE

Measures the average magnitude of the errors in the predictions. Lower values mean predictions are closer on average to actual data

MSE

Measures the average of the squared differences between forecast and observation. Lower values mean fewer and/or smaller large errors in predictions

Continuous Ranked Probability Score (CRPS)

Evaluate how well the predicted cumulative distribution function (CDF) matches the observed outcome. The prediction is modeled as a log-normal distribution for this calculation. The CRPS generalizes the Mean Absolute Error (MAE) to probabilistic forecasts. Lower CRPS values indicate better predictive performance.

Log Score

Measures the log-probability assigned to the observed value by the forecast distribution. The prediction is modeled as a log-normal distribution for this calculation. Larger log score values indicate better predictive performance.

Interval Score

Evaluates prediction intervals (e.g., 90% confidence intervals) of the forecast. The interval score penalizes forecasts when the observed value falls outside the predicted interval and favors narrower intervals when the observation is within the range. Lower values indicate better predictive performance.

Weighted Interval Score (WIS)

The WIS is a metric that combines multiple interval scores and serves as an approximation of the CRPS when forecasts are provided in the form of quantiles. Lower values indicate better predictive performance.