Hi, I'm trying to transform a PMML file to Graphml, but I've had some problems.

In the PMML file there are several child elements, but the one I want to work with is: AssociationModel and their children.

<?xml version="1.0" encoding="UTF-8"?>
<PMML version="3.0" xmlns="http://www.dmg.org/PMML-3_0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Header copyright="Copyright (c) Integral Solutions Ltd., 1994 - 2005. All rights reserved.">
</Header>
<DataDictionary numberOfFields="2">
<DataField name="cardid" optype="continuous" dataType="integer">
<Extension name="storageType" value="numeric"/>
</DataField>
<DataField name="Product" optype="categorical" dataType="string">
<Extension name="storageType" value="string"/>
<Value value="beer" property="valid"/>
<Value value="cannedmeat" property="valid"/>
<Value value="cannedveg" property="valid"/>
<Value value="confectionery" property="valid"/>
<Value value="dairy" property="valid"/>
<Value value="fish" property="valid"/>
<Value value="freshmeat" property="valid"/>
<Value value="frozenmeal" property="valid"/>
<Value value="fruitveg" property="valid"/>
<Value value="softdrink" property="valid"/>
<Value value="wine" property="valid"/>
</DataField>
</DataDictionary>
<AssociationModel modelName="SHOPPING_ASSOC" algorithmName="Carma" functionName="associationRules" numberOfTransactions="939" minimumSupport="0.177848775292865" minimumConfidence="0.32013201320132" numberOfItems="7" numberOfItemsets="10" numberOfRules="18">
<MiningSchema>
<MiningField name="cardid" usageType="group"/>
<MiningField name="Product" usageType="active"/>
</MiningSchema>
<Item id="5" value="wine"/>
<Item id="7" value="fruitveg"/>
<Item id="4" value="confectionery"/>
<Item id="3" value="frozenmeal"/>
<Item id="6" value="fish"/>
<Item id="2" value="cannedveg"/>
<Item id="1" value="beer"/>
<Itemset id="9" numberOfItems="1" support="0.31096912">
<ItemRef itemRef="6"/>
</Itemset>
<AssociationRule id="1" support="0.15548456" confidence="0.874251497005988" antecedent="1" consequent="2" lift="2.718285283737161"/>
</AssociationModel>
</PMML>

The next code is the stylesheet and the problem is that I can't get the xslt processor reads the AssociationModel only and then with the template process the child elements. Any suggestion? Thanks in advance.

?xml version="1.0" encoding="UTF-8"?>
<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform" version="1.0">

    <xsl:output method="xml"/>

    <xsl:template match="/">
        <xsl:comment>This GraphML was generated from a PMML</xsl:comment>
        <graphml xmlns="http://graphml.graphdrawing.org/xmlns">
            <graph edgedefault="undirected">
                <xsl:apply-templates select="AssociationModel"/>
            </graph>
        </graphml>
    </xsl:template>

    <xsl:template match="PMML">
        <xsl:comment>read pmml tag</xsl:comment>
             <xsl:apply-templates select="AssociationModel"/>
    </xsl:template>

    <xsl:template match="AssociationModel">
        <xsl:comment>read association_model tag</xsl:comment>
        <xsl:apply-templates/>
    </xsl:template>

    <xsl:template match="Item">
       <xsl:element name="node">
                <xsl:attribute name="id">
                    <xsl:comment>read Item tag</xsl:comment>
                    <xsl:value-of select="id"/>
                </xsl:attribute>
       </xsl:element>
    </xsl:template>

    <xsl:template match="Itemset">
        <xsl:apply-templates/>
    </xsl:template>

    <xsl:template match="AssociationRule">
    </xsl:template>

</xsl:stylesheet>

See
http://www.dpawson.co.uk/xsl/sect2/N5536.html
especially numbers 13 and 23.

--

Joe (Microsoft MVP - XML)

Your source document is in a namespace:

<PMML version="3.0" xmlns="http://www.dmg.org/PMML-3_0"

so you need to qualify all names with a prefix:

<xsl:stylesheet
  xmlns:pmml="http://www.dmg.org/PMML-3_0"

<xsl:template match="pmml:PMML">
  ...

In XSLT 2.0 you can say xpath-default-namespace="http://www.dmg.org/PMML-3_0"

on the xsl:stylesheet element (or elsewhere)

Michael Kay
http://www.saxonica.com/
Author, XSLT Programmer's Reference and XPath 2.0 Programmer's Reference

Thanks for the advice Joefawcett ;)

It works now, thanks mhkay!

Hi,

If you're interested I've done a little fun work with PMML. Mostly as a means to learn a bit of VB.Net and test PMML exported by SPSS Clementine. Some of it include using XSLT to parse association models and convert into HTML. See the links below.

KDKeys Data Mining with Clementine;
http://www.kdkeys.net/forums/thread/6495.aspx

SourceForge PMML homepage;
http://sourceforge.net/forum/forum.php?forum_id=187860

Cheers

Tim

HI, I need to transform a PMML file to SVG format but I´dont know how, please help me.....

<?xml version="1.0" ?>
  <PMML version="3.1" xmlns="http://www.dmg.org/PMML-3_1" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
    <Header copyright="dmg.org"/>
    <DataDictionary numberOfFields="3">
      <DataField name="marital status" optype="categorical" dataType="string">
          <Value value="s"/>
          <Value value="d"/>
          <Value value="m"/>
      </DataField>
      <DataField name="age" optype="continuous" dataType="double"/>
      <DataField name="salary" optype="continuous" dataType="double"/>
    </DataDictionary>
    <ClusteringModel modelName="Mini Clustering"
          functionName="clustering"
          modelClass="centerBased"
          numberOfClusters="2">
      <MiningSchema>
               <MiningField name="marital status"/>
               <MiningField name="age"/>
               <MiningField name="salary"/>
      </MiningSchema>

      <ComparisonMeasure kind="distance">
               <squaredEuclidean/>
      </ComparisonMeasure>

      <ClusteringField field="marital status"
                      compareFunction="absDiff"/>
      <ClusteringField field="age"
                      compareFunction="absDiff"/>
      <ClusteringField field="salary"
                      compareFunction="absDiff"/>
      <CenterFields>
        <DerivedField name="c1" optype="continuous" dataType="double">
          <NormContinuous field="age">
               <LinearNorm orig="45" norm="0"/>
               <LinearNorm orig="82" norm="0.5"/>
               <LinearNorm orig="105" norm="1"/>
          </NormContinuous>
        </DerivedField>
        <DerivedField name="c2" optype="continuous" dataType="double">
          <NormContinuous field="salary">
               <LinearNorm orig="39000" norm="0"/>
               <LinearNorm orig="39800" norm="0.5"/>
               <LinearNorm orig="41000" norm="1"/>
          </NormContinuous>
        </DerivedField>
        <DerivedField name="c3" optype="continuous" dataType="double">
          <NormDiscrete field="marital status" value="m"/>
        </DerivedField>
        <DerivedField name="c4" optype="continuous" dataType="double">
          <NormDiscrete field="marital status" value="d"/>
        </DerivedField>
        <DerivedField name="c5" optype="continuous" dataType="double">
          <NormDiscrete field="marital status" value="s"/>
        </DerivedField>
      </CenterFields>
      <MissingValueWeights>
        <Array n="5" type="real">1 1 1 1 1</Array>
      </MissingValueWeights>
      <Cluster name="marital status is d or s">
        <Array n="5" type="real">
               0.524561 0.486321 0.128427 0.459188 0.412384</Array>
      </Cluster>
      <Cluster name="marital status is m">
        <Array n="5" type="real">
               0.69946 0.419037 0.591226 0.173521 0.235253</Array>
      </Cluster>
    </ClusteringModel>
  </PMML>