Dear friends:
I need to use Box-Cox transformation to make my samples approximately normal.
For the moment, i do not care about inflation of confidence limits
in further estimates and consider the transformation as a stand-alone self
contained problem. I need to use the maximum likelihood approach, however I am
also curious about other possible approaches. Is there an economical way to
fulfill the task? My understading is that the function "boxcox" in the MASS
library is not exactly what I need. Prof. Ripley, please excuse me if I am
wrong and don't hesitate to correct me. I would appreciate any
ideas/codes/references. Thank you so much and have a great day
Stanley Rubinstein
Get 250 color business cards for FREE!
http://businesscards.lycos.com/vp/fastpath/
|