Hi guys :)
1. Could anyone suggest how to implement and use Pearson coefficient as distance measure?
2. Is it possible to extract only those variables that are used in weka learners, e.g. J48 decision tree? For example if I have 500 variables, the weka learner select only 100 of them. I want to extract the names of those 100 variables.
Regards
Hi,
I guess (and only guess) you could open your workspace and search your workflow for your J48 node then find the port1-> object->portoject.zip unzip it and find the mapper.xmlout file.
Parse the file with some xml knime node.
This file should begin with something like that :
<?xml version="1.0" encoding="UTF-8"?>
<config xmlns="http://www.knime.org/2008/09/XMLConfig" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.knime.org/2008/09/XMLConfig http://www.knime.org/XMLConfig_2008_09.xsd" key="model_info">
<config key="mapper">
<entry key="unique_index" type="xint" value="0"/>
<config key="origstringToString">
<entry key="Col10" type="xstring" value="Col10"/>
<entry key="Col20" type="xstring" value="Col20"/>
<entry key="Col12" type="xstring" value="Col12"/>
<entry key="Col11" type="xstring" value="Col11"/>
<entry key="Col14" type="xstring" value="Col14"/>
<entry key="Col13" type="xstring" value="Col13"/>
<entry key="Col16" type="xstring" value="Col16"/>
<entry key="Col8" type="xstring" value="Col8"/>
<entry key="Col15" type="xstring" value="Col15"/>
<entry key="Col9" type="xstring" value="Col9"/>
<entry key="Col18" type="xstring" value="Col18"/>
<entry key="Col17" type="xstring" value="Col17"/>
<entry key="Col4" type="xstring" value="Col4"/>
<entry key="Col19" type="xstring" value="Col19"/>
<entry key="Col5" type="xstring" value="Col5"/>
<entry key="Col6" type="xstring" value="Col6"/>
<entry key="Col7" type="xstring" value="Col7"/>
<entry key="Col0" type="xstring" value="Col0"/>
<entry key="Col1" type="xstring" value="Col1"/>
<entry key="Col2" type="xstring" value="Col2"/>
<entry key="Col3" type="xstring" value="Col3"/>
Best regards,
Fabien