This function calculates the percent prediction error of a sample based on the predicted and true values
Usage
ppe(true, pred, abs = TRUE)
Arguments
true
a numeric vector representing the true value on which a prediction (pred) will be compared
pred
a numeric vector with the predicted values, must be equal length to true
abs
a logical value indicating whether to return the absolute values for ppe, default is abs = TRUE
Value
Function returns a list of results, including the a list of all the precent prediction errors, along with the mean, 95 percent confidence intervals, range, and standard deviation.
Details
This function calculates the percent prediction error (ppe) as a scaled residual (Smith 1980), generally based on the absolute residual, so:
$$ppe=\frac{\left ( \left | true-pred \right | \right )}{pred}\times 100$$
References
Smith, R. J. (1980) Rethinking allometry. Journal of Theoretical Biology, 87, 97-111.