Computes the mean relative error by normalizing with the given values

metric_mean_relative_error(..., normalizer, name = NULL, dtype = NULL)

Arguments

...

Passed on to the underlying metric. Used for forwards and backwards compatibility.

normalizer

The normalizer values with same shape as predictions.

name

(Optional) string name of the metric instance.

dtype

(Optional) data type of the metric result.

Value

A (subclassed) Metric instance that can be passed directly to compile(metrics = ), or used as a standalone object. See ?Metric for example usage.

Details

This metric creates two local variables, total and count that are used to compute the mean relative error. This is weighted by sample_weight, and it is ultimately returned as mean_relative_error: an idempotent operation that simply divides total by count.

If sample_weight is NULL, weights default to 1. Use sample_weight of 0 to mask values.

metric = mean(|y_pred - y_true| / normalizer)

For example:

m = metric_mean_relative_error(normalizer=c(1, 3, 2, 3))
m$update_state(c(1, 3, 2, 3), c(2, 4, 6, 8))
 # result     = mean(c(1, 1, 4, 5) / c(1, 3, 2, 3)) = mean(c(1, 1/3, 2, 5/3))
 #            = 5/4 = 1.25
m$result()

See also