Range-to-RMSE ratio

Range-to-RMSE ratio (RER) scales RMSE by the observed range.

Usage

rer(obs, pred, na.rm = TRUE)

Arguments

obs

Numeric observation vector.

pred

Numeric prediction vector paired with obs.

na.rm

Logical; remove incomplete pairs?

Value

One numeric value.

Details

$$\mathrm{RER}=\frac{\max_{i}(obs_i)-\min_{i}(obs_i)} {\sqrt{n^{-1}\sum_{i=1}^{n}(obs_i-pred_i)^2}}.$$

RER is non-negative and larger values indicate smaller error relative to the observed range. It is sensitive to extreme observations. For perfect predictions with a nonzero range it returns Inf; it returns NA with a warning for the indeterminate zero-over-zero case. Missing-value handling follows bias().

References

Bellon-Maurel et al. (2010). See rpd().

See also

Other prediction metrics: mape(), mdae(), mpe(), msle(), rae(), rmsle(), rpd(), rpiq(), rrmse(), sep(), smape(), willmott_d()