Re: [R] Several Regression by combinations variables

2011-05-30 Thread Bert Gunter
In other words, "Abandon hope, all ye who enter..." -- Bert (Sorry, couldn't resist. Berlioz's "March to the Gallows" from Symphonie Fantastique also played in my head when I read Frank's comment, but I couldn't put that in the email.) On Mon, May 30, 2011 at 5:52 AM, Frank Harrell wrote: >

Re: [R] Several Regression by combinations variables

2011-05-30 Thread Frank Harrell
This is not accepted as good statistical practice, and the resulting inferential quantities will violate all known statistical principles. Frank Mohamed Lajnef wrote: > > Dear All, > I have 11 variables, and i would like generate combinations of those > variables by fours variables i,e 330 po

[R] Several Regression by combinations variables

2011-05-30 Thread Mohamed Lajnef
Dear All, I have 11 variables, and i would like generate combinations of those variables by fours variables i,e 330 possibility (choose(11,4)).? After that, make a regression analysis with this 330 possibility ? is there a program ( or package) to do that ? Any help would be appreciated Rega