How to reading clusters ?

Hi,

For our trainign data clustering models show the maximum accuracy.

We want to preset the results to our client. In order to do that we want to present what are the properties of each cluster.

For example for my predictable attribute if Cluster 9 is showing maximum population then I want to show what conditions of attributes make cluster 9 .

What is the way of doing this ?

Thanks,

Vkas

There are a couple of ways. You can get the cluster description from the content by issuing a query like

SELECT NODE_DESCRIPTION FROM MyModel.CONTENT WHERE NODE_CAPTION='Cluster 9'

The node description is a description of what attributes bring a case in to a cluster. You can also use the cluster discrimination pane of the viewer to show Cluster 9 vs all other clusters to see what is important for that cluster. You can copy the contents of that view into excel and edit for presentation (copy will be enhanced in SP2 coming the first part of 2007).

Additionally, the discrimination view is availalbe as a thin client (web) sample control that ships with the product. You will need to install the product samples to access this control.

|||

Hi,

Can I create reports on mining model using Reporting Services ? Could you pl. provide me some steps for that ?

thanks,

Vikas