f <- compf(h,e,fh,fe), REAL symmetric matrices h and e, positive integers fh and fe; integer vector INS must be defined |

NOTE: This macro is OBSOLETE and is retained only for backward compatibility because it was in file MacAnova.mac in earlier versions of MacAnova. For doing stepwise variable selection in discriminant analysis you should use newer macros dastepsetup(), daentervar(), daremovevar(), dastepstatus() and dasteplook(). Macro compf() computes Fs-to enter at any stage in stepwise variable selection in linear discriminant analysis. You can use compf() after manova() has computed hypothesis and error matrices H and E, with fh and fe degrees of freedom respectively. Integer vector INS must be defined, containing the variable numbers that are "in". INS = 0 means no variables are in and the Fs-to-enter are simply the separate ANOVA Fs. When INS != 0, the Fs-to-enter are the analysis of covariance Fs for each "out" variable, with the "in" variables being used as covariates. compf(H,E,fh,fe) returns structure(f:f_to_enter,df:vector(fh,fe-k), ins:INS, outs:OUTS), where OUTS is run(p) when INS = 0, is NULL when INS contains all integers 1, ..., p and is run(p)[-J] otherwise, where p = ncols(H). k is the number of variables "in". The F-to-enter statistics have nominal degrees of freedom fh and fe - k. Here is an example of starting forward stepwise variable selection. Cmd> manova("y = groups",silent:T)#, response matrix y, factor groups Cmd> H <- matrix(SS[2,,]); E <- matrix(SS[3,,]) Cmd> fh <- DF[2]; fe <- DF[3] Cmd> INS <- 0 # no variables in Cmd> results <- compf(H,E,fh,fe) Cmd> j <- grade(results$f,down:T)[1] # index of largest F Cmd> # now continue with forstep(j,H,E,fh,fe)

Gary Oehlert 2003-01-15