Vector Linear Regression - Multi Label

Oprick

Hello,
I'm trying to built a vector linear regression model with multiple label attributes, but I'm finding difficulties with the special role setup.

I read the vector linear regression description and from the below I understand that all labels must be named in its special role as something like "label_1", "label_2" and so on. 

****The attributes forming the vector should be marked as special, the special role names of all label attributes should start with 'label'.*****

I'm sure that this must be easy, but whatever I've tried doesn't work.

Enclosed the example process and the data

Can you please shed some light on this?

Thanks

TobiasMalbrecht

Hi @Oprick,

Try setting the role of the first label just to "label", the remainder ones to "label_2", "label_3", and so on. See attached process. This should work.

Hope this helps,

Tobias

sgenzer

hi @Oprick no you need to only label one target at a time.

<?xml version="1.0" encoding="UTF-8"?><process version="9.5.001">
  <context>
    <input/>
    <output/>
    <macros/>
  </context>
  <operator activated="true" class="process" compatibility="9.5.001" 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="read_excel" compatibility="9.5.001" expanded="true" height="68" name="Read Excel" width="90" x="45" y="34">
        <parameter key="excel_file" value="example.xlsx"/>
        <parameter key="sheet_selection" value="sheet number"/>
        <parameter key="sheet_number" value="1"/>
        <parameter key="imported_cell_range" value="A1"/>
        <parameter key="encoding" value="SYSTEM"/>
        <parameter key="first_row_as_names" value="true"/>
        <list key="annotations"/>
        <parameter key="date_format" value=""/>
        <parameter key="time_zone" value="SYSTEM"/>
        <parameter key="locale" value="English (United States)"/>
        <parameter key="read_all_values_as_polynominal" value="false"/>
        <list key="data_set_meta_data_information">
          <parameter key="0" value="A.true.integer.attribute"/>
          <parameter key="1" value="B.true.integer.attribute"/>
          <parameter key="2" value="C.true.integer.attribute"/>
          <parameter key="3" value="D.true.integer.attribute"/>
          <parameter key="4" value="E.true.integer.attribute"/>
          <parameter key="5" value="year.true.integer.attribute"/>
          <parameter key="6" value="week.true.integer.attribute"/>
          <parameter key="7" value="air.true.real.attribute"/>
          <parameter key="8" value="mmp.true.real.attribute"/>
          <parameter key="9" value="total.true.integer.attribute"/>
          <parameter key="10" value="total+1.true.integer.attribute"/>
          <parameter key="11" value="label_A+1.true.integer.attribute"/>
          <parameter key="12" value="label_B+1.true.integer.attribute"/>
          <parameter key="13" value="label_C+1.true.integer.attribute"/>
          <parameter key="14" value="label_D+1.true.integer.attribute"/>
          <parameter key="15" value="label_E+1.true.integer.attribute"/>
          <parameter key="16" value="total-1.true.integer.attribute"/>
          <parameter key="17" value="total-2.true.integer.attribute"/>
          <parameter key="18" value="total-3.true.integer.attribute"/>
          <parameter key="19" value="A-1.true.integer.attribute"/>
          <parameter key="20" value="A-2.true.integer.attribute"/>
          <parameter key="21" value="A-3.true.integer.attribute"/>
          <parameter key="22" value="B-1.true.integer.attribute"/>
          <parameter key="23" value="B-2.true.integer.attribute"/>
          <parameter key="24" value="B-3.true.integer.attribute"/>
          <parameter key="25" value="C-1.true.integer.attribute"/>
          <parameter key="26" value="C-2.true.integer.attribute"/>
          <parameter key="27" value="C-3.true.integer.attribute"/>
          <parameter key="28" value="D-1.true.integer.attribute"/>
          <parameter key="29" value="D-2.true.integer.attribute"/>
          <parameter key="30" value="D-3.true.integer.attribute"/>
          <parameter key="31" value="E-1.true.integer.attribute"/>
          <parameter key="32" value="E-2.true.integer.attribute"/>
          <parameter key="33" value="E-3.true.integer.attribute"/>
          <parameter key="34" value="id.true.integer.attribute"/>
        </list>
        <parameter key="read_not_matching_values_as_missings" value="false"/>
        <parameter key="datamanagement" value="double_array"/>
        <parameter key="data_management" value="auto"/>
      </operator>
      <operator activated="true" class="set_role" compatibility="9.5.001" expanded="true" height="82" name="Set Role" width="90" x="179" y="34">
        <parameter key="attribute_name" value="label_A+1"/>
        <parameter key="target_role" value="label"/>
        <list key="set_additional_roles"/>
      </operator>
      <operator activated="true" class="vector_linear_regression" compatibility="9.5.001" expanded="true" height="82" name="Vector Linear Regression" width="90" x="313" y="34">
        <parameter key="use_bias" value="true"/>
        <parameter key="ridge" value="1.0E-8"/>
      </operator>
      <operator activated="true" class="apply_model" compatibility="9.5.001" expanded="true" height="82" name="Apply Model" width="90" x="447" y="34">
        <list key="application_parameters"/>
        <parameter key="create_view" value="false"/>
      </operator>
      <connect from_op="Read Excel" from_port="output" to_op="Set Role" to_port="example set input"/>
      <connect from_op="Set Role" from_port="example set output" to_op="Vector Linear Regression" to_port="training set"/>
      <connect from_op="Vector Linear Regression" from_port="model" to_op="Apply Model" to_port="model"/>
      <connect from_op="Vector Linear Regression" from_port="exampleSet" to_op="Apply Model" to_port="unlabelled data"/>
      <connect from_op="Apply Model" from_port="labelled data" to_port="result 2"/>
      <connect from_op="Apply Model" from_port="model" to_port="result 1"/>
      <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>

Scott

Oprick

Hi @sgenzer,
when in VLR operator's description says (see below) I have to confess that the plural on word label lead me to believe that this operator was capable of dealing with multi-label problems.

Can you please clarify that?

"It regresses all regular attributes upon a vector of labels. The attributes forming the vector should be marked as special, the special role names of all label attributes should start with 'label'."

Thanks,
Pedro

sgenzer

hi @Oprick in all honesty I have never used that operator in 6+ years of using the software. Never even seen it before. Looking at the tutorial it seems to function like the other learners but I do see what you're saying. I just opened up the source code on GitHub and it looks like our good friend @TobiasMalbrecht was the author back in the day. Maybe he has some insight...

If you want multi-label modeling, I'd just use the multi-labeling modeling operator in the Ensembles folder. Open the tutorial and you will see how it works on Titanic.



Scott

Oprick

Hi @sgenzer, thanks for you help and pulling the Tobias into the thread.
I'm aware of multi-label indeed, but I have to say that I'm particular curious of this supposed capabilities of VLR.

Again thanks for your help,
Pedro

sgenzer

fair enough @Oprick. You will have to wait for the maestro @TobiasMalbrecht to get back from vacation

Oprick

Hello
 @sgenzer and @TobiasMalbrecht thank you very much to both of you.

Regards