Metrics

Regression Metrics

For deterministic forecasts (point predictions with num_samples == 1), probabilistic forecasts (num_samples > 1), and quantile forecasts. For probabilistic and quantile forecasts, use parameter q to define the quantile(s) to compute the deterministic metrics on:

  • Aggregated over time:
    Absolute metrics:

    • MERR: Mean Error

    • MAE: Mean Absolute Error

    • MSE: Mean Squared Error

    • RMSE: Root Mean Squared Error

    • RMSLE: Root Mean Squared Log Error

    Relative metrics:

    • MASE: Mean Absolute Scaled Error

    • MSSE: Mean Squared Scaled Error

    • RMSSE: Root Mean Squared Scaled Error

    • MAPE: Mean Absolute Percentage Error

    • wMAPE: weighted Mean Absolute Percentage Error

    • sMAPE: symmetric Mean Absolute Percentage Error

    • OPE: Overall Percentage Error

    • MARRE: Mean Absolute Ranged Relative Error

    Other metrics:

    • R2: Coefficient of Determination

    • CV: Coefficient of Variation

  • Per time step:
    Absolute metrics:

    • ERR: Error

    • AE: Absolute Error

    • SE: Squared Error

    • SLE: Squared Log Error

    Relative metrics:

    • ASE: Absolute Scaled Error

    • SSE: Squared Scaled Error

    • APE: Absolute Percentage Error

    • sAPE: symmetric Absolute Percentage Error

    • ARRE: Absolute Ranged Relative Error

For probabilistic forecasts (storchastic predictions with num_samples >> 1) and quantile forecasts:

  • Aggregated over time:
    Quantile metrics:

    • MQL: Mean Quantile Loss

    • QR: Quantile Risk

    Quantile interval metrics:

    • MIW: Mean Interval Width

    • MWS: Mean Interval Winkler Score

    • MIC: Mean Interval Coverage

    • MINCS_QR: Mean Interval Non-Conformity Score for Quantile Regression

  • Per time step:
    Quantile metrics:

    • QL: Quantile Loss

    Quantile interval metrics:

    • IW: Interval Width

    • WS: Interval Winkler Score

    • IC: Interval Coverage

    • INCS_QR: Interval Non-Conformity Score for Quantile Regression

Classification Metrics

For deterministic forecasts (point predictions with num_samples == 1), probabilistic forecasts (num_samples > 1), and class label probability forecasts. Most metrics allow to extract the scores for specific labels with parameter labels.

Dynamic Time Warping Metrics

For Dynamic Time Warping (DTW) (aggregated over time):

  • DTW: Dynamic Time Warping Metric

previous

<no title>

next

Metrics