gamlss (version 5.2-0)

devianceIncr: The global deviance increment

Description

The global deviance increment is the contribution of each individual observation to the global deviance. The function devianceIncr() can be used to extract the global deviance increment for a fitted gamlss model or for a new (test/validation) data set. Large values for global deviance increment indicate a bad fit and for new data a bad prediction.

Usage

devianceIncr(obj, newdata = NULL)

Arguments

obj

a gamlss object

newdata

test data set to check the global deviance increment.

Value

Returns a vector of the global deviance increments for each observation.

References

Rigby, R. A. and Stasinopoulos D. M. (2005). Generalized additive models for location, scale and shape,(with discussion), Appl. Statist., 54, part 3, 1-38.

Rigby, R. A., Stasinopoulos, D. M., Heller, G. Z., and De Bastiani, F. (2019) Distributions for modeling location, scale, and shape: Using GAMLSS in R, Chapman and Hall/CRC. An older version can be found in https://www.gamlss.com/.

Stasinopoulos D. M. Rigby R.A. (2007) Generalized additive models for location scale and shape (GAMLSS) in R. Journal of Statistical Software, Vol. 23, Issue 7, Dec 2007, https://www.jstatsoft.org/v23/i07/.

Stasinopoulos D. M., Rigby R.A., Heller G., Voudouris V., and De Bastiani F., (2017) Flexible Regression and Smoothing: Using GAMLSS in R, Chapman and Hall/CRC.

(see also https://www.gamlss.com/).

See Also

deviance

Examples

Run this code
# NOT RUN {
#-----------------------------------------------------------------
# Count data set
# fit Poisson model
h1 <- gamlss(Claims~L_Population+L_Accidents+L_KI+L_Popdensity, 
             data=LGAclaims, family=PO)
p1<-devianceIncr(h1)
# fit negative binomial model
h2 <- gamlss(Claims~L_Population+L_Accidents+L_KI+L_Popdensity, 
             data=LGAclaims, family=NBI)
p2<-devianceIncr(h2)
# comparing using boxplots
boxplot(cbind(p1,p2))
# comparing using emphirical cdf
plot(ecdf(p1))
lines(ecdf(p2), col=2)
# comparing agaist the y-values
plot(p1~LGAclaims$Claims, pch=20, col="gray")
points(p2~LGAclaims$Claims, pch="-", col="orange")
#----------------------------------------------------------------
# Continuous data sets
# }
# NOT RUN {
m1 <- gamlss(head~pb(age), data=db[1:6000,])
p1<-devianceIncr(m1)
m2 <- gamlss(head~pb(age), sigma.fo=~pb(age), nu.fo=~pb(age), 
      tau.fo=~pb(age), data=db[1:6000,], family=BCT)
p2<-d.evianceIncr(m2)
# comparing using summaries
summary(p1); summary(p2)
# comparing using boxplots
boxplot(cbind(p1,p2))
# comparing using histograms
hist(p1, col=rgb(1,0,0,0.5), xlim=c(0,50), breaks=seq(0,50,2))
hist(p2, col=rgb(0,0,1,0.5), add=T)
# comparing using emphirical cdf
plot(ecdf(p1))
lines(ecdf(p2), col=2)
# }
# NOT RUN {
#----------------------------------------------------------------
# }

Run the code above in your browser using DataCamp Workspace