topic Miner decision tree error message in SAS Data Science

Miner decision tree error message

rogelio_mancisidor — Wed, 07 Oct 2015 20:07:52 GMT

Hello,

I have some decision trees in my miner project that were working ok. I mean, i build them and then i set 'Use Frozen Tree' to 'Yes'.. However, after a recent update to the project's Imput data source (I added new variables), some of the Decision Tree nodes crashed, but not all of them. Here is the 1st error message in one of the Tree node that crashed:

ERROR: PROC internal error. Routine called at inappropriate moment.

I have taken a look to the table Tree_emtree in the Tree folder from one of the Tree node that crashed and compared it with the same table from a Tree node that did not crashed, and I cannot see any difference in those tables. Then I tried to use those 2 tables to replicate the decion trees, but using new nodes. I can only replicate the trees that didnt crashed.

does any one have an idea what can be happening? why not all of the trees crashed but only some of them? the option 'Use Frozen Tree' did not work?

Re: Miner decision tree error message

M_Maldonado — Thu, 22 Oct 2015 20:39:56 GMT

Hi Rogelio,

I could not reproduce this behavior. Did you get this to work?

This is what I did and it went OK.

1. I created a data source Mydata_A, and ran a decision tree.

2. Next I used base SAS to add 5 more columns to Mydata_A. Then I went to Enterprise Miner to drag-and-drop the data source again. The advisor took care of the roles and levels of the extra columns, but to do this the proper way I could have refreshed the metadata. You can also delete this data source, and create it again.

3. I connected a decision tree node, and specified Import Tree Model as Yes, and selected the tree set (in this example EMWS1.tree_emtree).

4. Run

Is this what didn't work for you?

Thanks,

Miguel

Re: Miner decision tree error message

mathew17 — Tue, 27 Oct 2015 13:03:03 GMT

The Decision Tree Algorithm is implemented with reasonable defaults for splitting and termination criteria. It is unlikely that you will need to use any of the build settings that are supported for Decision Tree. The settings are described as follows.

Settings to specify the homogeneity metric for finding the optimal split condition:

TREE_IMPURITY_METRIC can be either gini or entropy. The default is gini.

Settings to control the growth of the tree:

TREE_TERM_MAX_DEPTH specifies the maximum depth of the tree, from root to leaf inclusive. The default is 7.
TREE_TERM_MINPCT_MODE specifies the minimum number of cases required in a child node, expressed as a percentage of the rows in the training data. The default is .05%.
TREE_TERM_MINPCT_SPLIT specifies the minimum number of cases required in a node in order for a further split to be possible. Expressed as a percentage of all the rows in the training data. The default is 1%.
TREE_TERM_MINREC_MODE specifies the minimum number of cases required in a child node. Default is 10.
TREE_TERM_MINREC_SPLIT specifies the minimum number of cases required in a node in order for a further split to be possible. Default is 20. Java training in chennai