topic Scoring New Data On Enterprise Miner (Clustering Node) in SAS Data Science

Scoring New Data On Enterprise Miner (Clustering Node)

omerzeybek — Fri, 11 Sep 2015 01:49:57 GMT

I have already clustered my customers by using SAS Miner using a specific month's data. I want to implement produced rules to the historical data set to understand consistency of clustering.

However, when I use the "Flow Code" of Clustering Node, historical dataset is grouped in only one cluster. How can I implement selected clustering rules to the new dataset?

Re: Scoring New Data On Enterprise Miner (Clustering Node)

M_Maldonado — Fri, 11 Sep 2015 05:02:33 GMT

How similar are the training and the scoring data sets? Do both have the same input variables? Are the distributions similar for the input variables across both data sets?

Try looking at the score code (if it is not too large) and identify why all your observations go to the same cluster.