Pages: [1]
  Print  
Author Topic: re-use of previous calculation (intermediate result)  (Read 970 times)
VD
Guest
« on: March 20, 2009, 02:06:40 PM »

Hello,

in my model I would like to iterate through the maximum number of generations. Does rapidminer re-use the  results from the previous step in order to save calculation time? I mean after finishing the calculation with only one feature (and saving this result to file, or something similar) it can directly continue to add the next attributes to the best n-features and continue instead of starting again from 1 for the next iteration ...


<operator name="Root" class="Process" expanded="yes">
    <parameter key="logfile"   value="logfile.log"/>
    <parameter key="resultfile"   value="result.res"/>
    <operator name="ExcelExampleSource" class="ExcelExampleSource">
        <parameter key="excel_file"   value="Data.xls"/>
        <parameter key="first_row_as_names"   value="true"/>
        <parameter key="create_label"   value="true"/>
    </operator>
    <operator name="ParameterIteration" class="ParameterIteration" expanded="yes">
        <list key="parameters">
          <parameter key="FeatureSelection.maximum_number_of_generations"   value="1,2,3,4"/>
        </list>
        <operator name="FeatureSelection" class="FeatureSelection" expanded="yes">
            <parameter key="show_population_plotter"   value="true"/>
            <parameter key="plot_generations"   value="1"/>
            <parameter key="keep_best"   value="10"/>
            <parameter key="maximum_number_of_generations"   value="3"/>
            <operator name="LinearRegression" class="LinearRegression">
                <parameter key="keep_example_set"   value="true"/>
                <parameter key="feature_selection"   value="none"/>
            </operator>
            <operator name="ModelApplier" class="ModelApplier">
                <parameter key="keep_model"   value="true"/>
                <list key="application_parameters">
                </list>
                <parameter key="create_view"   value="true"/>
            </operator>
            <operator name="RegressionPerformance" class="RegressionPerformance">
                <parameter key="keep_example_set"   value="true"/>
                <parameter key="main_criterion"   value="squared_correlation"/>
                <parameter key="correlation"   value="true"/>
                <parameter key="squared_correlation"   value="true"/>
                <parameter key="skip_undefined_labels"   value="false"/>
                <parameter key="use_example_weights"   value="false"/>
            </operator>
        </operator>
    </operator>
</operator>

Thanks and Best Regards

Volker
Logged
VD
Guest
« Reply #1 on: March 23, 2009, 09:34:58 AM »

Hello,

... or is it possible / how is it possible to save the best feature combination (for each generation) in a separate file ?

Best Regards

Volker
Logged
Ingo Mierswa
Administrator
Hero Member
*****
Posts: 1221



WWW
« Reply #2 on: March 23, 2009, 12:48:23 PM »

Hi Volker,

both requests are only possible for the evolutionary feature selection and weighting schemes. There are parameters for "intermediate" weights savings as well as initilizing the operator by input weights. But this is not possible for the greedy feature selection heuristics.

Cheers,
Ingo
Logged

Did you try our new Marketplace? Upload or download new Extensions, add comments, and organize your operators. Have a look at  http://marketplace.rapid-i.com
Pages: [1]
  Print  
 
Jump to: