fel.repeated {hysteresis}

Methods for easily fitting multiple ellipses from repeated measures designs.

Description

Fit a sinusoidal hysteretic process between an input and an output variable across multiple loops separated by subjects and repeated.

Usage

fel.repeated(x,y=NULL,subjects=NULL,repeated=NULL,subjects.in="all",repeated.in="all",...)
## S3 method for class 'ellipsefitlist'
summary(object,N=1000,boot=TRUE,seed=NULL,...)

Arguments

x

numeric input vector.

y

numeric output vector.

subjects

factor of the same length as x that represents experimental units.

repeated

factor of the same length as x that represents the repeated measure.

subjects.in

a vector of characters, the levels of subjects to be included. Default is "all".

repeated.in

a vector of characters, the levels of repeated to be included. Default is "all".

object

an ellipsefitlist object.

N

Number of bootstrap replicates.

boot

whether to use bootstrapping to obtain standard errors and less biased parameter estimates.

seed

for generating random numbers. See summary.fittedloop.

...

extra arguments to either fel or summary.ellipsefitlist.

Details

Fits multiple ellipses with one call, separated by the factors subjects and repeated. The arguments subjects.in and repeated.in are used to select subsets of the factors subjects and repeated.

Value

fel.repeated returns an object of class ellipsefitlist.

models

Separate model fits for each ellipse, see fel.

Estimates

Parameter estimates for all ellipses in matrix form.

Std.Errors

Delta standard errors for all ellipses in matrix form.

When boot=TRUE fel.repeated returns an object of class ellipsesummarylist which consists of

models

a vector of separate model summaries for each ellipse, see summary.ellipsefit.

values

Bootstrapped parameter estimates, standard errors, quantiles, and more for each ellipse.

Boot.Estimates

Bootstrapped parameter estimates with reduced bias.

Boot.Std.Errors

Standard errors provided by bootstrapping.

Author(s)

Spencer Maynes, Fan Yang, and Anne Parkhurst.

References

Yang, F. and A. Parkhurst, Efficient Estimation of Elliptical Hysteresis (submitted)

See Also

fel for a more general way to fit multiple ellipses, or for fitting just one ellipse. plot.ellipsefit for plotting and summary.ellipsefit for summarizing and bootstrapping an ellipsefitlist object. Also residuals.ellipsefitlist.

Examples

## Select 2 subjects with 2 replications and fit 4 ellipses
data(EllipseData) 
emodels.rep <- fel.repeated(EllipseData$X, EllipseData$Y, method = "harmonic2", 
subjects = EllipseData$subjects,subjects.in=c("A","C"),
repeated=EllipseData$repeated)
emodels.rep                #Gives estimates and delta standard errors 
emodels.rep$Estimates      #List estimates only
emodels.rep$Std.Errors     #List delta standard errors 
par(mfrow=c(2,2))
plot(emodels.rep, main="Repeated Ellipses",xlab="X",ylab="Y")
par(mfrow=c(1,1))

### Bootstrap estimates and standard errors (Seed is necessary if want to reproduce results)
boot.rep.ellipse<-fel.repeated(EllipseData$X,EllipseData$Y,method = "harmonic2",
subjects = EllipseData$subjects,subjects.in=c("A","C"),
repeated=EllipseData$repeated,boot=TRUE,seed=123)
boot.rep.ellipse  #Gives boot estimates, boot bias, boot SE and boot quartiles
par(mfrow=c(2,2))
plot(boot.rep.ellipse, main="Repeated Ellipses",xlab="X",ylab="Y",values="ellipse")
par(mfrow=c(1,1))

##Can write results to a file. First set your directory from the file tab.
#Change file path in command below to coincide with where you want to store data files
#setwd("C:/Users................")
#write.table(boot.rep.ellipse$Boot.Estimates,"Ellipes.eg.repbootvalues.txt")
#test.fel=read.table("Ellipes.eg.repbootvalues.txt",header=TRUE)
#head(test.fel)