merge from qigong branch

[qigongFDA]

1.Create a R project to generate R package for iMRMC and iRoeMetz. 2.Comment out sizing validation code to speed up software.

[brandon-gallas]

Please explain to me why sizing validation is commented out. What is the speed difference? Why do we not need this code? Will this code be uncommented after the package is created? Should we separate the sizing validation from the main work flow used in the package?

[qigongFDA]

1. Sizing validation code do 8 sizing studies (1 or 3 split plot; paired or unpaired reader and case) for each trail simulation. For one trial we do 1(data) + 8 (sizing) = 9 analysis. I didn't test the speed, but I think it should slow down the code.
2. We add these codes for validate sizing function. It is for development. For end user, they don't need those results.
3. Current release version doesn't comment out these code, I think we should do comment these in next version. We only uncomment them when we need to do sizing validation in future.
4. Now we do sizing analysis analysis right after we finish one trail analysis. As we don't need to do validation very often, I don't think we should separate it.

[brandon-gallas]

I don't like this. If we need the sizing validation elements, we will do one of two things:

• Test the code and see if there is a speed issue.
  • If there is a speed issue, create another java file, maybe it could be called RMvalidate.java, or we will create a flag to control whether or not we do the extra sizing evaluations.