SAS Enterprise Miner - How cluster node deal with missing values?

[carol_kvarg]

Hi,

Im working with a dataset which contains at least 20% of missing values. I intentionally didn't use the Input node. My problem is that I noticed in result of clustering node frequency of data points within the cluster is as the dataset would not contain missing values.

I couldn't find straight answer, if clustering node input missing values as a default, if so, what is the input method is used?

Does it count missing values for the frequency?

Moreover, is there any documentation on whether cluster node have predefined data dimensionality reduction technique?

[tom_grant]

I just did a small test & for my dataset observations with missing values were removed from cluster results.