Interpretation Extension: retrieve model from repository

anaRodrigues

Hello,

I would like to generate a SHAP interpretation for a model I have stored in the repository. Is this possible? It doesn't seem to work.

<?xml version="1.0" encoding="UTF-8"?><process version="9.9.000">
  <context>
    <input/>
    <output/>
    <macros/>
  </context>
  <operator activated="true" class="process" compatibility="9.9.000" expanded="true" name="Process">
    <parameter key="logverbosity" value="init"/>
    <parameter key="random_seed" value="2001"/>
    <parameter key="send_mail" value="never"/>
    <parameter key="notification_email" value=""/>
    <parameter key="process_duration_for_mail" value="30"/>
    <parameter key="encoding" value="SYSTEM"/>
    <process expanded="true">
      <operator activated="true" class="retrieve" compatibility="9.9.000" expanded="true" height="68" name="Retrieve G_D_SVM-RFE_DT" width="90" x="246" y="34">
        <parameter key="repository_entry" value="//Local Repository/Models_SVM-RFE/G_D_SVM-RFE_DT"/>
      </operator>
      <operator activated="true" class="retrieve" compatibility="9.9.000" expanded="true" height="68" name="Retrieve gland_trainSet_stable" width="90" x="112" y="136">
        <parameter key="repository_entry" value="//Local Repository/gland_trainSet_stable"/>
      </operator>
      <operator activated="true" class="multiply" compatibility="9.9.000" expanded="true" height="103" name="Multiply" width="90" x="246" y="136"/>
      <operator activated="true" class="interpretation:generate_interpretation" compatibility="0.1.001" expanded="true" height="124" name="Generate Interpretation" width="90" x="447" y="85">
        <parameter key="algorithm" value="Shapley"/>
        <parameter key="sample_size" value="100"/>
        <parameter key="redraw_local_samples" value="true"/>
        <parameter key="explanation_algorithm" value="Correlation"/>
        <parameter key="locality" value="0.2"/>
        <parameter key="use_local_random_seed" value="false"/>
        <parameter key="local_random_seed" value="1992"/>
      </operator>
      <connect from_op="Retrieve G_D_SVM-RFE_DT" from_port="output" to_op="Generate Interpretation" to_port="mod"/>
      <connect from_op="Retrieve gland_trainSet_stable" from_port="output" to_op="Multiply" to_port="input"/>
      <connect from_op="Multiply" from_port="output 1" to_op="Generate Interpretation" to_port="training"/>
      <connect from_op="Multiply" from_port="output 2" to_op="Generate Interpretation" to_port="test"/>
      <connect from_op="Generate Interpretation" from_port="importance" to_port="result 1"/>
      <connect from_op="Generate Interpretation" from_port="global weights" to_port="result 2"/>
      <portSpacing port="source_input 1" spacing="0"/>
      <portSpacing port="sink_result 1" spacing="0"/>
      <portSpacing port="sink_result 2" spacing="0"/>
      <portSpacing port="sink_result 3" spacing="0"/>
    </process>
  </operator>
</process>

Thanks in advance,
Ana

anaRodrigues

Hi Martin,

All regular attributes are type 'real' as they should be. I think the problem is that the model was trained with an example set that went through feature selection, so the sets of attributes are not the same. I didn't think this would be an issue because the 'apply model' operator works just fine when I input the test set with the full set of attributes.

Do you know of any way to fix this?

Thanks,
Ana

MartinLiebig

Hi,

whats the error message?

Best,

Martin

anaRodrigues

Hi Martin,

Here it is.


Thank you,
Ana

MartinLiebig

thats not a problem on the model side. The data you want to have explained has a different type in application compare to training of the model.

Can you check the type? Likely it moved to nominal but is a numerical?

Best,

Martin

MartinLiebig

Hi,

i can only tell you what the operator shows you. And this is that this certain attribute is different to the training set. Usually a superset of your attributes should work well.

For more details I would need to see the model and the exampleset.

BR,

Martin