topic cluster data graphical representation in Statistical Procedures

cluster data graphical representation

michtka — Thu, 03 May 2012 21:39:28 GMT

Hi to everyone, checking this link from SAS about the rpoc mixel model:

http://support.sas.com/documentation/cdl/en/statug/63033/HTML/default/viewer.htm#statug_mixed_sect006.htm

I got the next example:

Clustered Data Example

Consider the following SAS data set as an introductory example:

   data heights; input Family Gender$ Height @@; datalines; 1 F 67   1 F 66   1 F 64   1 M 71   1 M 72   2 F 63 2 F 63   2 F 67   2 M 69   2 M 68   2 M 70   3 F 63 3 M 64   4 F 67   4 F 66   4 M 67   4 M 67   4 M 69 ;

The response variable Height measures the heights (in inches) of 18 individuals. The individuals are classified according to Family and Gender. You can perform a traditional two-way analysis of variance of these data with the following PROC MIXED statements:

My question is, there are some procedure to visualize graphically this kind of clustering problems, i.e to see which model is fitting better

my raw of data, or see how well separated the cluter of data are, etc, etc.

Thnaks in advance,

Re: cluster data graphical representation

PGStats — Fri, 04 May 2012 02:13:19 GMT

You can visualize the group distributions of observed and residual values from this model with :

proc mixed data=heights plots=boxplot(fixed observed);

class Family Gender;

model Height = Gender Family Family*Gender;

run;

A kind word of caution: you can learn a lot from SAS documentation, but it cannot replace a course in statistics.

Re: cluster data graphical representation

michtka — Fri, 04 May 2012 11:46:26 GMT

Thanks PG...I found like I was looking for:

http://support.sas.com/documentation/cdl/en/statugclustering/61759/PDF/default/statugclustering.pdf