GoFCompare: Goodness of fit comparison (single sample)

In UKFE: UK Flood Estimation

Goodness of fit comparison (single sample)

Description

compares the RMSE of four distribution fits for a single AMAX sample.

Usage

GoFCompare(x)

Arguments

x	a numeric vector (your AMAX sample)

Details

This function calculates an RMSE fit score for four distributions (GEV, GenLog, Gumbel, & Kappa3). The lowest RMSE is the best fit. It works as follows. For each distribution: Step1. Simulate 500 samples the same size as x. Step2. Calculate the mean across all 500 samples for each rank to create an ordered central estimate. Step3. Calculate the RMSE between the result of step 2 and the ordered x. Step4. Standardise the RMSE by dividing it by the mean of x and multiply it by 100 (RMSE as a percentage of mean). Note that this is not a hypothesis test. It is only for comparing the fit across the distributions.

Value

A list. The first element is a dataframe with four columns and one row of results. Each column has the standardised RMSE associated with one of the four distributions (GEV, GenLog, Gumbel, Kappa3). The second element is a character string stating the distribution with the best fit.

Author(s)

Anthony Hammond

Examples

# Get an AMAX sample then compare the fit..
AM15006 <- GetAM(15006)
GoFCompare(AM15006$Flow)

UKFE documentation built on Sept. 30, 2024, 9:46 a.m.