Demonstration of the command "descr" of S-Plus for 
discrimination and classification. 

** Data set: Table 11.1 of the textbook (Riding-mower owners)

> dim(T11.1)  % Data was loaded before.
[1] 24  3
> T11.1[1,]
  income  lot grp 
1     60 18.4   1
> x=T11.1[,1:2]

> plot(x[,1],x[,2],xlab='Income',ylab='lotsize')
> text(x[,1],x[,2],T11.1[,3])

> m1=discr(x,2)
> m1
$cor:
[1] 7.338967e-001 3.020311e-007

$vars:
           [,1]        [,2] 
[1,] 0.04953562  0.03870816
[2,] 0.38805246 -0.32448712

$groups:
           [,1]      [,2] 
[1,]  0.7071066 0.7071070
[2,] -0.7071070 0.7071066

> y=as.matrix(x)%*%m1$vars   % Obtain the discriminant variables
> plot(y[,1],y[,2],xlab='first var',ylab='second var')
> text(y[,1],y[,2],T11.1[,3])

*** Iris Data set Exercise 11.28, Table 11.5
> dim(T11.5)
[1] 150   5
> y1=T11.5[,1]/T11.5[,2]
> y2=T11.5[,3]/T11.5[,4]

> y=cbind(log(y1),log(y2))
> plot(y[,1],y[,2],xlab='sepal',ylab='petal')
> text(y[,1],y[,2],T11.5[,5])

> m2=discr(y,3)
> m2
$cor:
[1] 0.9397170 0.1202879 0.0000000

$vars:
          [,1]      [,2] 
[1,]  8.872250 -5.820336
[2,] -2.927041 -2.993714

$groups:
           [,1]        [,2]      [,3] 
[1,] -0.8114433  0.09069956 0.5773504
[2,]  0.3271736 -0.74808037 0.5773503
[3,]  0.4842699  0.65738076 0.5773501

> z=y%*%m2$vars
> plot(z[,1],z[,2],xlab='first var',ylab='second var')
> text(z[,1],z[,2],T11.5[,5])
>