0

I'm trying to create hypothetical parameters to run the following simulation in r;

library(nlme)

ITS.single.group = function(y, time, time.intrv1, time.intrv2, pchi, smallmodel, greatmodel){
  
  # y is response; 
  # time is observation time; 
  # time.intrv1+1 is the time point of the first intervention;
  # time.intrv2+1 is the time point of the second intervention
  # pchi is the critical value rejecting the hypothesis
  # *************************************************************************
  # great model involves the small model,  
  # for example greatmodel = formula(y ~ time + x1 + x2 + time1:x1 + time2:x2) 
  # and smallmodel = formula(y~time)
  # *************************************************************************
  
  nsmp = length(y)
  
  x1 = c(rep(0,time.intrv1), rep(1,nsmp-time.intrv1))
  x2 = c(rep(0,time.intrv2), rep(1,nsmp-time.intrv2))
  
  time1 = time-time.intrv1-1
  time2 = time-time.intrv2-1
  
  fit.small = gls(smallmodel, correlation=corAR1(form=~1), method = "ML") 
  # fit small model
  
  fit.great = gls(greatmodel, correlation=corAR1(form=~1), method = "ML") 
  # fit great model
  
  dif = fit.great$logLik - fit.small$logLik
  rej = (2*dif > pchi)  
  rej
}

Can anyone help with hypothetical parameters that I can use to run the model?

4

0 回答 0