topic Re: HPSPLIT node discription in SAS Data Science

HPSPLIT node discription

akoohi — Thu, 16 Mar 2023 16:44:15 GMT

Hello,

I am interested in exact node association in the tree generated by hpsplit. For example, of the leaf node is 45 what is its parent node in the tree? Is there any dataset that I can look at as an output of the HPSPLIT tree that has this information?

Thanks,

Arezou

Re: HPSPLIT node discription

ballardw — Thu, 16 Mar 2023 18:09:03 GMT

You may get a more useful example by providing data and or using or one of the examples provided by SAS with code similar to what you are using.

Re: HPSPLIT node discription

akoohi — Thu, 16 Mar 2023 19:42:16 GMT

Here is the output of HPSPLIT tree, is there any way to get an output option that shows the parent of node A is node 7, for all of the nodes in datasets?

Re: HPSPLIT node discription

ballardw — Sun, 19 Mar 2023 07:19:36 GMT

Using one of the SAS supplied examples for HPSPLIT to start with:

ods graphics on;

proc hpsplit data=sashelp.baseball seed=123;
   class league division;
   model logSalary = nAtBat nHits nHome nRuns nRBI nBB
                     yrMajor crAtBat crHits crHome crRuns crRbi
                     crBB league division nOuts nAssts nError;
   output out=hpsplout;
   ods output wholetreeplot= work.wholetree
              zoomedtreeplot= work.zoomtree
   ;
run;

Looking at the Details section, ODS Graphics, of the online help for Proc Hpsplit the plot names for nodes are Wholetreeplot and ZoomedTreePlot. Each of these contains parent/child node information though a bit differently. The nodes labeled A are using hex notation instead of 10, B for 11 and so on. So you may need to think through the display a bit.

Re: HPSPLIT node discription

akoohi — Thu, 16 Mar 2023 21:13:49 GMT

omg, you are a life saver. thank you so so much.