ny_default_load_train.xml

<?xml version="1.0" ?>
<Simulation verbosity="debug">
  <!-- NOTE: THE DATA FOR THIS INPUT FILE IS NOT INCLUDED IN THIS REPOSITORY!
   THIS IS JUST INCLUDED FOR POSTERITY SO IT CAN BE TRACKED UNTIL IT'S MOVED TO FORCE -->
  <RunInfo>
    <WorkingDir>Output</WorkingDir>
    <Sequence>load, train, meta, serialize, sample</Sequence>
  </RunInfo>

  <Files>
    <Input name="input">Data.csv</Input>
    <Input name="pk">arma_202112_nyiso_def.pk</Input>
  </Files>

  <Steps>
    <IOStep name="load">
      <Input class="Files" type="">input</Input>
      <Output class="DataObjects" type="HistorySet">input</Output>
    </IOStep>

    <RomTrainer name="train">
      <Input class="DataObjects" type="HistorySet">input</Input>
      <Output class="Models" type="ROM">arma</Output>
    </RomTrainer>

    <IOStep name="meta">
      <Input class="Models" type="ROM">arma</Input>
      <Output class="DataObjects" type="DataSet">meta</Output>
      <Output class="OutStreams" type="Print">romMeta</Output>
    </IOStep>

    <IOStep name="serialize">
      <Input class="Models" type="ROM">arma</Input>
      <Output class="Files" type="">pk</Output>
    </IOStep>

    <MultiRun name="sample">
      <Input class="DataObjects" type="PointSet">placeholder</Input>
      <Model class="Models" type="ROM">arma</Model>
      <Sampler class="Samplers" type="MonteCarlo">mc</Sampler>
      <Output class="DataObjects" type="DataSet">synth</Output>
      <Output class="OutStreams" type="Print">synth</Output>
    </MultiRun>

  </Steps>

  <DataObjects>
    <PointSet name="placeholder">
      <Input>scaling</Input>
      <Output>OutputPlaceHolder</Output>
    </PointSet>

    <HistorySet name="input">
      <Input>scaling,YEAR</Input>
      <Output>TOTALLOAD,WIND,SOLAR,HOUR</Output>
      <options>
        <pivotParameter>HOUR</pivotParameter>
      </options>
    </HistorySet>

    <DataSet name="synth">
      <Input>scaling</Input>
      <Output>TOTALLOAD,WIND,SOLAR</Output>
      <Index var="YEAR">TOTALLOAD,WIND,SOLAR</Index>
      <Index var="HOUR">TOTALLOAD,WIND,SOLAR</Index>
    </DataSet>
    <DataSet name="meta"/>
  </DataObjects>

  <Models>
    <ROM name="arma" subType="ARMA">
      <pivotParameter>HOUR</pivotParameter>
      <Features>scaling</Features>
      <Target>TOTALLOAD,WIND,SOLAR,HOUR</Target>
      <P>1</P>
      <Q>0</Q>
      <Fourier>24, 12</Fourier>
      <SpecificFourier variables='TOTALLOAD,WIND'>
        <periods>8760, 4380, 2920, 2190, 438, 168, 24, 12, 6, 3</periods>
      </SpecificFourier>
      <ZeroFilter>SOLAR</ZeroFilter>
      <outTruncation domain='positive'>SOLAR</outTruncation>
      <preserveInputCDF>True</preserveInputCDF>
      <reseedCopies>True</reseedCopies>
      <seed>42</seed>
      <Segment grouping="interpolate">
        <macroParameter>YEAR</macroParameter>
        <Classifier class="Models" type="PostProcessor">classifier</Classifier>
        <evalMode>full</evalMode>
        <subspace pivotLength="24" shift="zero">HOUR</subspace>
        <evaluationClusterChoice>random</evaluationClusterChoice>
      </Segment>
    </ROM>
    <PostProcessor name="classifier" subType="DataMining">
      <KDD labelFeature="labels" lib="SciKitLearn">
        <Features>TOTALLOAD</Features>
        <SKLtype>cluster|KMeans</SKLtype>
        <n_clusters>2</n_clusters>
        <tol>1E-12</tol>
        <init>k-means++</init>
        <random_state>3</random_state>
        <precompute_distances>True</precompute_distances>
      </KDD>
    </PostProcessor>
  </Models>

  <OutStreams>
    <Print name="romMeta">
      <type>csv</type>
      <source>meta</source>
    </Print>

    <Print name="synth">
      <type>csv</type>
      <source>synth</source>
    </Print>
  </OutStreams>

  <Samplers>
    <MonteCarlo name="mc">
      <samplerInit>
        <limit>1</limit>
        <initialSeed>42</initialSeed>
      </samplerInit>
      <constant name="scaling">1.0</constant>
    </MonteCarlo>
  </Samplers>
</Simulation>