PMML in Action: Unleashing the Power of Open Standards for Data Mining and Predictive Analytics by Wen-Ching Lin & Tridivesh Jena & Alex Guazzelli

Author:Wen-Ching Lin & Tridivesh Jena & Alex Guazzelli
Language: eng
Format: mobi
Publisher: CreateSpace
Published: 2010-01-30T16:00:00+00:00

Finally, a cluster is represented by the Cluster element. For a center-based model, each cluster is defined in terms of its center coordinates which is represented by an Array element. If the model is distribution-based, an array is still used, however it no longer represents the center coordinates of the cluster.

If the clustering algorithm used was a Kohonen map, the cluster position in a 3-dimensional map may be represented in the KohonenMap element. This element is used for informational purposes only.

The PMML code shown in Figure 7.2 implements the definition of a clustering model including the distance measures as well as the cluster definitions for a distance-based model. Note that in this example, two clusters are defined, named "X" and "Y". For cluster "X", the center coordinates for fields "var0", "var1", and "var2" are "0.121", "0.424", and "0.234", respectively.

<ComparisonMeasure kind="distance">

<euclidean/>

</ComparisonMeasure>

<ClusteringField field="var0" compareFunction="delta"/>

<ClusteringField field="var1" compareFunction="delta"/>

<ClusteringField field="var2" compareFunction="equal"/>

<Cluster id="1" name="X">

<Array n="3" type="real">0.121 0.424 0.234</Array>

</Cluster>

<Cluster id="2" name="Y">

<Array n="3" type="real">0.908 0.982 0.259</Array>

</Cluster>

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.