/* Suggested solution for SAS 3 exercise */
/* Bjarne Kjær Ersbøll */

title 'SAS-exercise 3';

proc print data=sasdata.iris;
title2 'Print of data';

proc discrim data=sasdata.iris wcov wcorr pcov pcorr list pool=yes;
class species;
title2 'Linear discriminant analysis - all data';

proc discrim data=sasdata.iris wcov wcorr pcov pcorr listerr pool=yes;
class species;
priors setosa=0.01 versicolor=0.98 virginica=0.01;
title2 'Linear discriminant analysis - all data - new priors';

proc discrim data=sasdata.iris wcov wcorr pcov pcorr listerr pool=yes;
class species;
var sepallen petallen;
title2 'Linear discriminant analysis - all data - only 2 variables';

proc discrim data=sasdata.iris wcov wcorr pcov pcorr listerr pool=test;
class species;
title2 'Discriminant analysis - quadratic/linear determined by test';

proc print data=sasdata.iriscali;
title2 'Print of training data';

proc print data=sasdata.iristest;
title2 'Print of test data';


proc discrim data=sasdata.iriscali wcov wcorr pcov pcorr list pool=yes
    testdata=sasdata.iristest testlist;
  class species;
  testclass species;
title2 'Linear discriminant analysis - independent training and test data';

proc candisc data=stat2.iris out=toplot distance anova; 
class species;
title2 'Canonical discriminant analysis - all data';

* a plot of the canonical variables is interesting, so; 

proc plot data=toplot; 
plot can2*can1=species; 
title2 'Canonical discriminant analysis - all data';
title3 'plot on first 2 canonical variables';
run;