best_distribution: Distribution with Best Error Intervals

In errint: Builds Error Intervals

Description

best_distribution computes the distribution assumption that gives error intervals with the lower accuracy error for a given set of residuals.

Usage

best_distribution(phi, errors, dists = c("n", "nm", "l", "lm", "w", "b",
  "moge"), ...)

Arguments

phi	residual values used to compute the error interval.
errors	set of real errors corresponding to the predictions of a particular model.
dists	character vector with the distribution assumptions to test. See also 'Details'.
...	additional arguments to be passed to functions error_interval and acc_intervals.

Details

Allowed distribution assumptions are:

• "n": Zero-mu Gaussian

• "nm": General Gaussian

• "l": Zero-mu Laplace

• "lm": General Laplace

• "b": Beta

• "w": Weibull

• "moge": Moge

Value

Returns an object of class c("df_intervals", "data.frame") with information of the distribution assumption with lower accuracy error.

Author(s)

Jesus Prada, [e-mail]

References

Link to the scientific paper

Prada, Jesus, and Jose Ramon Dorronsoro. "SVRs and Uncertainty Estimates in Wind Energy Prediction." Advances in Computational Intelligence. Springer International Publishing, 2015. 564-577,

with theoretical background for this package is provided below.

http://link.springer.com/chapter/10.1007/978-3-319-19222-2_47

See Also

df_intervals error_interval acc_intervals

Examples

best_distribution(rnorm(10),rnorm(10),dists=c("n","b"))

errint documentation built on May 1, 2019, 10:06 p.m.