Alternatives to accuracy and bias metrics based on percentage errors for radiation belt modeling applications
- Los Alamos National Lab. (LANL), Los Alamos, NM (United States)
This report reviews existing literature describing forecast accuracy metrics, concentrating on those based on relative errors and percentage errors. We then review how the most common of these metrics, the mean absolute percentage error (MAPE), has been applied in recent radiation belt modeling literature. Finally, we describe metrics based on the ratios of predicted to observed values (the accuracy ratio) that address the drawbacks inherent in using MAPE. Specifically, we define and recommend the median log accuracy ratio as a measure of bias and the median symmetric accuracy as a measure of accuracy.
- Research Organization:
- Los Alamos National Laboratory (LANL), Los Alamos, NM (United States)
- Sponsoring Organization:
- USDOE National Nuclear Security Administration (NNSA)
- DOE Contract Number:
- AC52-06NA25396
- OSTI ID:
- 1260362
- Report Number(s):
- LA-UR-16-24592; TRN: US1601539
- Country of Publication:
- United States
- Language:
- English
Similar Records
Measures of model performance based on the log accuracy ratio
Accounting for Training Data Error in Machine Learning Applied to Earth Observations
Temporal ensemble learning of univariate methods for short term load forecasting
Journal Article
·
Wed Jan 03 00:00:00 EST 2018
· Space Weather
·
OSTI ID:1260362
Accounting for Training Data Error in Machine Learning Applied to Earth Observations
Journal Article
·
Mon Mar 23 00:00:00 EDT 2020
· Remote Sensing
·
OSTI ID:1260362
+17 more
Temporal ensemble learning of univariate methods for short term load forecasting
Conference
·
Tue Feb 20 00:00:00 EST 2018
·
OSTI ID:1260362