forsearch_nls {forsearch}

Create Statistics Of Forward Search in a Nonlinear Model Database

Description

Prepares summary statistics at each stage of forward search for subsequent plotting. Forward search is conducted in two steps: Step 1 to identify minimal set of observations to estimate unknown parameters, and Step 2 to add one observation at each stage such that observations in the set are best fitting at that stage.

Usage

forsearch_nls(phaselist, data, poolstart, poolformula, algorithm=
  "default", controlarg=NULL, initial.sample=1000, skip.step1=NULL, 
  begin.diagnose=100, verbose=TRUE)

Arguments

phaselist

LIST of formula, formulacont, start, nopl for each phase

data

Name of database. First 2 variables are Observation and Phases (both mandatory)

poolstart

List Start values for Step 2

poolformula

Formula for pooled data from all phases for Step 2

algorithm

algorithm for nls function.

controlarg

nls control. Default is NULL to use preset nls.control

initial.sample

Number of observation sets in Step 1 of forward search

skip.step1

NULL or a vector of integers for observations to be included in Step 1

begin.diagnose

Numeric. Indicates where in code to begin printing diagnostics. 0 prints all; 100 prints none

verbose

TRUE causes function identifier to display before and after run

Details

All datasets are considered to be in phases. See vignette for definition and discussion. There is a phaselist for each phase and an element for each phaselist input variable. In addition, there is a (pool)start and a (pool)formula input variable for the pooled dataset.

Value

LIST

Rows in stage

Observation numbers of rows included at each stage

Standardized residuals

Matrix of errors at each stage

Number of model parameters

Same as number of levels of poolstart input variable

Sigma

Estimate of random error at final stage; used to standardize all residuals

Fixed parameter estimates

Vector of parameter estimates at each stage

s^2

Estimate of random error at each stage

Call

Call to this function

Author(s)

William R. Fairweather

References

Atkinson, A and M Riani. Robust Diagnostic Regression Analysis, Springer, New York, 2000. Pinheiro, JC and DM Bates. Mixed Effects Models in S and S-PLUS, Springer, New York, 2000. Example from nlstools package

Examples

## Not run: 
t<-(0:35)/3
VO2<-c(377.1111,333.3333,352.1429,328.7500,369.8750,394.4000,352.6667,337.3333,
  366.4286,364.0000,293.8889,387.0000,364.8889,342.2222,400.3000,375.1111,
  320.5556,385.1667,527.0714,688.6364,890.8182,1145.1538,1254.9091,1327.5000,
  1463.9000,1487.8333,1586.6667,1619.1000,1494.4167,1640.4545,1643.3750,
  1583.6364,1610.8000,1568.5000,1464.5833,1652.8000)
Observation <- 1:36
Phases <- as.factor(c(rep("REST",18), rep("EXERCISE",18)))
test01 <- data.frame(Observation,Phases,t,VO2)

formula.1 <-as.formula(VO2~VO2rest)
formulacont.1 <- as.formula(VO2~VO2rest)
start.1 <- list(VO2rest = 400)
nopl.1 <- 1

formula.2<-
  as.formula(VO2~(VO2rest+(VO2peak-VO2rest)*(1-exp(-(t-5.883)*I(1/mu)))))
formulacont.2<-
  as.formula(VO2~(VO2rest+(VO2peak-VO2rest)*(1-exp(-(t-5.883)*I(1/mu)))))
start.2 <- list(VO2rest = 400, VO2peak = 1600, mu = 1)
nopl.2 <- 6

phaselist <- list(
             REST=
 list(formula=formula.1,formulacont=formulacont.1,start=start.1,nopp=nopl.1),
             EXERCISE=
 list(formula=formula.2,formulacont=formulacont.2,start=start.2,nopp=nopl.2))

pstart <- list(VO2rest=400, VO2peak = 1600, mu = 1)
pformula <- as.formula(VO2~(t<=5.883)*(VO2rest)+          
            (t>5.883)*(VO2rest+(VO2peak-VO2rest)*
            (1-exp(-(t-5.883)*I(1/mu)))))
forsearch_nls(phaselist=phaselist, data=test01, 
   poolstart=pstart, poolformula=pformula, algorithm="default", 
   controlarg=nls.control(maxiter=50,warnOnly=TRUE), initial.sample = 155)

## End(Not run)