Computes Root Mean Square Error
Computes the root mean square error for an array of forecasts, var_exp and
an array of observations, var_obs, which should have the same dimensions
except along the posloop dimension where the lengths can be different, with
the number of experiments/models for var_exp (nexp) and the number of
obserational datasets for var_obs (nobs).
The RMSE is computed along the posRMS dimension which should correspond to
the startdate dimension.
If compROW is given, the RMSE is computed only if rows along the compROW
dimension are complete between limits[1] and limits[2], i.e. there are no
NAs between limits[1] and limits[2]. This option can be activated if the
user wishes to account only for the forecasts for which observations are
available at all leadtimes.
Default: limits[1] = 1 and limits[2] = length(compROW dimension).
The confidence interval relies on a chi2 distribution.
.RMS provides the same functionality but taking a matrix of ensemble
members as input (exp).
RMS( var_exp, var_obs, posloop = 1, posRMS = 2, compROW = NULL, limits = NULL, siglev = 0.95, conf = TRUE ) .RMS(exp, obs, siglev = 0.95, conf = TRUE)
var_exp |
Matrix of experimental data. |
var_obs |
Matrix of observational data, same dimensions as var_exp except along posloop dimension, where the length can be nobs instead of nexp. |
posloop |
Dimension nobs and nexp. |
posRMS |
Dimension along which RMSE are to be computed (the dimension of the start dates). |
compROW |
Data taken into account only if (compROW)th row is complete. |
limits |
Complete between limits[1] & limits[2]. Default = NULL. |
siglev |
Confidence level of the computed confidence interval. 0.95 by default. |
conf |
Whether to compute confidence interval or not. TRUE by default. |
exp |
N by M matrix of N forecasts from M ensemble members. |
obs |
Vector of the corresponding observations of length N. |
RMS: Array with dimensions:
c(length(posloop) in var_exp, length(posloop) in var_obs, 1 or 3, all
other dimensions of var_exp & var_obs except posRMS).
The 3rd dimension corresponds to the lower limit of the 95% confidence
interval (only present if conf = TRUE), the RMSE, and the upper
limit of the 95% confidence interval (only present if
conf = TRUE).
.RMS:
$rms |
The root mean square error, |
$conf_low |
Corresponding to the lower limit of the |
$conf_high |
Corresponding to the upper limit of the |
History:
0.1 - 2011-05 (V. Guemas, vguemas@ic3.cat) - Original code
1.0 - 2013-09 (N. Manubens, nicolau.manubens2@ic3.cat) - Formatting to R CRAN
1.1 - 2017-02 (A. Hunter, alasdair.hunter@bsc.es) - Adapted to veriApply()
# Load sample data as in Load() example:
example(Load)
clim <- Clim(sampleData$mod, sampleData$obs)
ano_exp <- Ano(sampleData$mod, clim$clim_exp)
ano_obs <- Ano(sampleData$obs, clim$clim_obs)
runmean_months <- 12
dim_to_smooth <- 4 # Smooth along lead-times
smooth_ano_exp <- Smoothing(ano_exp, runmean_months, dim_to_smooth)
smooth_ano_obs <- Smoothing(ano_obs, runmean_months, dim_to_smooth)
dim_to_mean <- 2 # Mean along members
# Discard start-dates for which some leadtimes are missing
required_complete_row <- 3
leadtimes_per_startdate <- 60
rms <- RMS(Mean1Dim(smooth_ano_exp, dim_to_mean),
Mean1Dim(smooth_ano_obs, dim_to_mean),
compROW = required_complete_row,
limits = c(ceiling((runmean_months + 1) / 2),
leadtimes_per_startdate - floor(runmean_months / 2)))
PlotVsLTime(rms, toptitle = "Root Mean Square Error", ytitle = "K",
monini = 11, limits = NULL, listexp = c('CMIP5 IC3'),
listobs = c('ERSST'), biglab = FALSE, hlines = c(0),
fileout = 'tos_rms.eps')
# The following example uses veriApply combined with .RMS instead of RMS
## Not run:
require(easyVerification)
RMS2 <- s2dverification:::.RMS
rms2 <- veriApply("RMS2",
smooth_ano_exp,
# see ?veriApply for how to use the 'parallel' option
Mean1Dim(smooth_ano_obs, dim_to_mean),
tdim = 3, ensdim = 2)
## End(Not run)Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.