residuals.fittedloop {hysteresis}

Residuals, studentized residuals and fitted values for the hysteresis package.

Description

Extract input, output, geometric and algebraic residuals, studentized residuals and fitted values from fitted loops or ellipses.

Usage

##S3 methods for classes 'ellipsefit', 'ellipsesummary', 'fittedloop',
##'loopsummary','ellipsefitlist', 'ellipsesummarylist', 'fittedlooplist',
##'loopsummarylist', 'loop2r', 'fittedlooplist2r',
##'loopsummarylist2r' and 'loop2rsummary'.
## S3 method for class 'ellipsefit'
residuals(object,...)
## S3 method for class 'ellipsefit'
rstudent(model,...)
## S3 method for class 'ellipsefit'
fitted(object,...)

Arguments

object

an object created by fel or floop.

model

an object created by fel or floop.

...

other arguments.

Details

Geometric residuals are based on the straight line distance between predicted and true values along an x,y cartesian plane, and algebraic residuals are based on the method used to calculate the ellipsefit object. If method="harmonic2" (which is always the case if this is a fittedloop object) or if bootstrapping has occurred, then, there are no algebraic residuals and residuals.ellipsefit replaces these with the geometric residuals.

Studentization for the rstudent function is performed as if method="harmonic2" regardless of the method used for fitting the ellipse/loop. Therefore, unless method="harmonic2" and no bootstrapping is performed, these are pseudo-studentized residuals, not true studentized residuals. This is internal scaling studentization. Studentization for bootstrapping in the functions summary.ellipsefit/summary.fittedloop differs from the studentization performed by rstudent in that it only accounts for the influence matrix and does not divide by the standard deviation.

Value

input

a numeric vector. Observed input - fitted input for residuals.

output

a numeric vector. Observed output - fitted output for residuals.

geometric

a numeric vector. Not available with rstudent. See details.

algebraic

a numeric vector. Not available with rstudent or when the 'harmonic2' method is used. See details.

Author(s)

Spencer Maynes, Fan Yang, and Anne Parkhurst.

References

Yang, F. and A. Parkhurst, Estimating Elliptical Hysteresis: A Comparison of Analytic Methods. (submitted)

See Also

fel, floop, summary.ellipsefit and summary.fittedloop.

Examples

##For multiple loops/ellipses
data(HysteresisData)
Mloopmodels.rep <- floop.repeated(HysteresisData$X, HysteresisData$Y,
n=5,m=3, subjects = HysteresisData$subjects,subjects.in=c("A","C"),
repeated=HysteresisData$repeated)
Mloopmodels.rep                #Gives estimates and delta standard errors
residuals(Mloopmodels.rep)  #$input $output $geometric 
fitted(Mloopmodels.rep)     #$input $output
scatterplotMatrix(cbind(residuals(Mloopmodels.rep)$input,
residuals(Mloopmodels.rep)$output,residuals(Mloopmodels.rep)$geometric,
fitted(Mloopmodels.rep)$input,fitted(Mloopmodels.rep)$output),
  main='Residuals for Multiple Hysteresis Loops',smooth=FALSE,
  var.labels=c("Input Resid","Output Resid","Geometric Resid",
  "Fitted Input", "Fitted Output"),
  groups=residuals(Mloopmodels.rep)$repeated)
rstudent(Mloopmodels.rep)  #$input $output
scatterplotMatrix(cbind(rstudent(Mloopmodels.rep)$input,
rstudent(Mloopmodels.rep)$output,fitted(Mloopmodels.rep)$input,
fitted(Mloopmodels.rep)$output),main='Studentized Residuals 
for Multiple Hysteresis Loops',smooth=FALSE,
var.labels=c("Input Resid","Output Resid", "Fitted Input",
"Fitted Output"),groups=residuals(Mloopmodels.rep)$repeated)

##For single Ellipse
ellipse1 <- mel(sd.x=0.2,sd.y=0.04)
ellipse1.fit <- fel(ellipse1$x,ellipse1$y)
residuals(ellipse1.fit)
fitted(ellipse1.fit)
scatterplotMatrix(cbind(residuals(ellipse1.fit)$input,
residuals(ellipse1.fit)$output,residuals(ellipse1.fit)$geometric,
fitted(ellipse1.fit)$input,fitted(ellipse1.fit)$output),
main='Residuals for Simulated Ellipse',smooth=FALSE,
var.labels=c("Input Resid","Output Resid","Geometric Resid", 
"Fitted Input", "Fitted Output"))

rstudent(ellipse1.fit) #for input and output variables
scatterplotMatrix(cbind(rstudent(ellipse1.fit)$input,
rstudent(ellipse1.fit)$output,fitted(ellipse1.fit)$input,
fitted(ellipse1.fit)$output),main='Studentized Residuals
for Simulated Ellipse',smooth=FALSE,
var.labels=c("Input Resid","Output Resid","Fitted Input",
"Fitted Output"))

plot(ellipse1.fit$pred.y,rstudent(ellipse1.fit)$output,
xlab="Fitted Output",ylab="Output Studentized Residuals",
main="Studentized Residuals:Simulated Ellipse")
abline(h = 0, lty = 2, col = "gray")
qqnorm(rstudent(ellipse1.fit)$output,sub='Output Studentized
Residuals Simulated Ellipse')
qqline(rstudent(ellipse1.fit)$output,col="red") #q-q line