IA_tab: Indexes of Agreement Table

In nlraa: Nonlinear Regression for Agricultural Applications

Indexes of Agreement Table

Description

Indexes of agreement

printing function for IA_tab

plotting function for a IA_tab, it requires ‘ggplot2’

Usage

IA_tab(obs, sim, object, null.object)

## S3 method for class 'IA_tab'
print(x, ..., digits = 2)

## S3 method for class 'IA_tab'
plot(x, y, ..., type = c("OvsS", "RvsS"))

Arguments

obs	vector with observed data
sim	vector with simulated data (should be the same length as observed)
object	alternative to the previous two arguments. An object of class ‘lm’, ‘nls’, ‘lme’ or ‘nlme’
null.object	optional object which represents the ‘null’ model. It is an intercept-only model by default. (Not used at the moment).
x	object of class ‘IA_tab’.
...	additional plotting arguments (none use at the moment).
digits	number of digits for rounding (default is 2)
y	not used at the moment
type	either “OvsS” (observed vs. simulated) or “RvsS” (residuals vs. simulated).

Details

This function returns several indexes that might be useful for interpretation. Notice that bias (mean), rss, mse and rmse are model-free

The intercept, slope, reg.rss, reg.mse and reg.rmse are based on a regression model

For objects of class ‘lm’ and ‘nls’
bias: mean(obs - sim)
intercept: intercept of the model obs ~ beta_0 + beta_1 * sim + error
slope: slope of the model obs ~ beta_0 + beta_1 * sim + error
rss: residual sum of squares (model free)
mse: mean squared error (model free)
rmse: root mean squared error (model free)
reg. rss (deviance): residual sum of squares of the previous model
reg. mse (reg. rss / n): mean squared error; where n is the number of observations
reg. rmse: squared root of the previous index
R2.1: R-squared extracted from an ‘lm’ object
R2.2: R-squared computed as the correlation between observed and simulated to the power of 2.
ME: model efficiency
NME: Normalized model efficiency
Corr: correlation between observed and simulated
ConCorr: concordance correlation

For objects of class ‘lme’ or ‘nlme’ there is the marginal and conditional R2.

https://en.wikipedia.org/wiki/Coefficient_of_determination
https://en.wikipedia.org/wiki/Nash-Sutcliffe_model_efficiency_coefficient
https://en.wikipedia.org/wiki/Concordance_correlation_coefficient

See Also

IC_tab

Examples

require(nlme)
require(ggplot2)
## Fit a simple model and then compute IAs
data(swpg)
#' ## Linear model
fit0 <- lm(lfgr ~ ftsw + I(ftsw^2), data = swpg)
ias0 <- IA_tab(object = fit0)
ias0
## Nonlinear model
fit1 <- nls(lfgr ~ SSblin(ftsw, a, b, xs, c), data = swpg)
ias1 <- IA_tab(object = fit1)
ias1
plot(ias1)
## Linear Mixed Models
data(barley, package = "nlraa")
fit2 <- lme(yield ~ NF + I(NF^2), random = ~ 1 | year, data = barley)
ias2 <- IA_tab(object = fit2)
ias2
## Nonlinear Mixed Model
barleyG <- groupedData(yield ~ NF | year, data = barley)
fit3L <- nlsLMList(yield ~ SSquadp3(NF, a, b, c), data = barleyG)
fit3 <- nlme(fit3L, random = pdDiag(a + b ~ 1))
ias3 <- IA_tab(object = fit3)
ias3
plot(ias3)
## Plotting model
prds <- predict_nlme(fit3, interval = "conf", plevel = 0)
barleyGA <- cbind(barleyG, prds)
ggplot(data = barleyGA, aes(x = NF, y = yield)) + 
   geom_point() + 
   geom_line(aes(y = Estimate)) + 
   geom_ribbon(aes(ymin = Q2.5, ymax = Q97.5), 
               fill = "purple", alpha = 0.2)
## R2M for model 2
R2M(fit2)
## R2M for model 3
R2M(fit3)

## Using IA_tab without a model
IA_tab(obs = swpg$lfgr, sim = fitted(fit0))

nlraa documentation built on Aug. 21, 2025, 5:59 p.m.