Outputs - Revision history

Javdrher: /* Complex handling */

2014-02-28T08:52:04Z

Complex handling

← Older revision		Revision as of 08:52, 28 February 2014
Line 72:		Line 72:

	<Output name="S11,S12" complexHandling="real">		<Output name="S11,S12" complexHandling="real">
	<~~SampleSelector~~>lola</~~SampleSelector~~>		<SequentialDesign>lola</SequentialDesign>
	<Measure type="CrossValidation" target=".01" />		<Measure type="CrossValidation" target=".01" />
	</Output>		</Output>
Line 80:		Line 80:

	<Output name="S12" complexHandling="complex">		<Output name="S12" complexHandling="complex">
	<~~SampleSelector~~>error</~~SampleSelector~~>		<SequentialDesign>error</SequentialDesign>
	<~~AdaptiveModelBuilder~~>svmps</~~AdaptiveModelBuilder~~>		<ModelBuilder>svmps</ModelBuilder>
	<Measure type="ValidationSet" target=".05" />		<Measure type="ValidationSet" target=".05" />
	<Measure type="CrossValidation" target=".05" use="off" />		<Measure type="CrossValidation" target=".05" use="off" />

Javdrher: /* Component customization */

2014-02-28T08:51:39Z

Component customization

← Older revision		Revision as of 08:51, 28 February 2014
Line 37:		Line 37:
	<Outputs>		<Outputs>
	<Output name="out">		<Output name="out">
	<~~SampleSelector~~>lola</~~SampleSelector~~>		<SequentialDesign>lola</SequentialDesign>
	<Measure type="CrossValidation" target=".01" errorFcn="meanSquareError" use="on" />		<Measure type="CrossValidation" target=".01" errorFcn="meanSquareError" use="on" />
	<~~AdaptiveModelBuilder~~>lssvmps</~~AdaptiveModelBuilder~~>		<ModelBuilder>lssvmps</ModelBuilder>
	</Output>		</Output>

	<Output name="outinverse">		<Output name="outinverse">
	<~~SampleSelector~~>delaunay</~~SampleSelector~~>		<SequentialDesign>delaunay</SequentialDesign>
	<Measure type="CrossValidation" target=".05" errorFcn="maxRelativeError" use="on" />		<Measure type="CrossValidation" target=".05" errorFcn="maxRelativeError" use="on" />
	<Measure type="MinMax" />		<Measure type="MinMax" />
Line 50:		Line 50:
	</source>		</source>

	This configuration models the first Output named "out" using the CrossValidation [[Measures\|measure]] (using the meanSquareError function, and a target of 0.1) with the LOLA sample selector to select new samples. It builds models for this output using the lssvmps model builder. The second output, named "outinverse", uses CrossValidation in combination with the MinMax measure, and uses the delaunay sample selector to select new samples for this output. It builds models using the model builder defined higher up in the xml hierarchy. As you can see, you can add multiple measures to one Output, so that each measure has something to say in deciding the accuracy of the models. However, you can only select one ~~SampleSelector~~ or ~~AdaptiveModelBuilder~~.		This configuration models the first Output named "out" using the CrossValidation [[Measures\|measure]] (using the meanSquareError function, and a target of 0.1) with the LOLA sample selector to select new samples. It builds models for this output using the lssvmps model builder. The second output, named "outinverse", uses CrossValidation in combination with the MinMax measure, and uses the delaunay sample selector to select new samples for this output. It builds models using the model builder defined higher up in the xml hierarchy. As you can see, you can add multiple measures to one Output, so that each measure has something to say in deciding the accuracy of the models. However, you can only select one SequentialDesign or odelBuilder.

	For information on how multiple measures are handled and the configuration options of each measure, see [[Measures]].		For information on how multiple measures are handled and the configuration options of each measure, see [[Measures]].

Mnguyen: /* Default behavior, all outputs modeled */

2012-01-16T14:48:11Z

Default behavior, all outputs modeled

← Older revision		Revision as of 14:48, 16 January 2012
Line 3:		Line 3:
	== Default behavior, all outputs modeled ==		== Default behavior, all outputs modeled ==

	If no Outputs tag is defined in the configuration file, all outputs are modeled and evaluated with the ~~default CrossValidation~~ [[Measures\|measure]] ~~and target accuracy 0.01~~. Complex outputs are modeled directly. The default error function is rootRelativeSquareError.m.		If no Outputs tag is defined in the configuration file, all outputs are modeled and evaluated with the [[Measures\|measure]] specified by the plan or the run. Complex outputs are modeled directly. The default error function is rootRelativeSquareError.m.

	== Default behavior, selected outputs modeled ==		== Default behavior, selected outputs modeled ==

Dgorissen: /* Complex handling */

2009-09-29T18:14:08Z

Complex handling

← Older revision		Revision as of 18:14, 29 September 2009
Line 67:		Line 67:

	<source lang="xml">		<source lang="xml">
	<!-- Model outputs S11 and S12 using the ~~gradient~~ sample selector (overwrites the beforementioned setting).		<!-- Model outputs S11 and S12 using the LOLA sample selector (overwrites the beforementioned setting).
	Uses cross-validation for model evaluation. Both complex outputs are treated as two separate real numbers		Uses cross-validation for model evaluation. Both complex outputs are treated as two separate real numbers
	with no relation to each other, resulting in total in 4 reals that will be modeled for these two complex outputs. -->		with no relation to each other, resulting in total in 4 reals that will be modeled for these two complex outputs. -->

	<Output name="S11,S12" complexHandling="real">		<Output name="S11,S12" complexHandling="real">
	<SampleSelector>~~gradient~~</SampleSelector>		<SampleSelector>lola</SampleSelector>
	<Measure type="CrossValidation" target=".01" />		<Measure type="CrossValidation" target=".01" />
	</Output>		</Output>

	<!-- Model output S12 as a complex number using the ~~grid~~ sample selector. Models are evaluated using the ValidationSet measure.		<!-- Model output S12 as a complex number using the error sample selector. Models are evaluated using the ValidationSet measure.
	The CrossValidation accuracy of the model is also calculated for comparison purposes, but isn't used in the modeling algorithm. -->		The CrossValidation accuracy of the model is also calculated for comparison purposes, but isn't used in the modeling algorithm. -->

Dgorissen: /* Default behavior, all outputs modeled */

2009-09-29T18:13:01Z

Default behavior, all outputs modeled

← Older revision		Revision as of 18:13, 29 September 2009
Line 3:		Line 3:
	== Default behavior, all outputs modeled ==		== Default behavior, all outputs modeled ==

	If no Outputs tag is defined in the configuration file, all outputs are modeled and evaluated with the default CrossValidation [[Measures\|measure]] and target accuracy 0.01. Complex outputs are modeled directly.		If no Outputs tag is defined in the configuration file, all outputs are modeled and evaluated with the default CrossValidation [[Measures\|measure]] and target accuracy 0.01. Complex outputs are modeled directly. The default error function is rootRelativeSquareError.m.

	== Default behavior, selected outputs modeled ==		== Default behavior, selected outputs modeled ==

Dgorissen: /* Component customization */

2009-09-29T18:12:37Z

Component customization

← Older revision		Revision as of 18:12, 29 September 2009
Line 37:		Line 37:
	<Outputs>		<Outputs>
	<Output name="out">		<Output name="out">
	<SampleSelector>~~gradient~~</SampleSelector>		<SampleSelector>lola</SampleSelector>
	<Measure type="CrossValidation" target=".~~0001~~" use="on" />		<Measure type="CrossValidation" target=".01" errorFcn="meanSquareError" use="on" />
	<AdaptiveModelBuilder>~~svm~~</AdaptiveModelBuilder>		<AdaptiveModelBuilder>lssvmps</AdaptiveModelBuilder>
	</Output>		</Output>

	<Output name="outinverse">		<Output name="outinverse">
	<SampleSelector>~~grid~~</SampleSelector>		<SampleSelector>delaunay</SampleSelector>
	<Measure type="CrossValidation" target=".05" use="on" />		<Measure type="CrossValidation" target=".05" errorFcn="maxRelativeError" use="on" />
	<Measure type="MinMax" />		<Measure type="MinMax" />
	</Output>		</Output>
Line 50:		Line 50:
	</source>		</source>

	This configuration models the first Output named "out" using the CrossValidation measure, ~~but~~ with ~~increased target accuracy, and uses~~ the ~~gradient~~ sample selector to select new samples. It builds models for this output using the ~~svm~~ model builder. The second output, named "outinverse", uses CrossValidation in combination with the MinMax measure, and uses the ~~grid~~ sample selector to select new samples for this output. It builds models using the model builder defined higher up in the xml hierarchy. As you can see, you can add multiple measures to one Output, so that each measure has something to say in deciding the accuracy of the models. However, you can only select one SampleSelector or AdaptiveModelBuilder.		This configuration models the first Output named "out" using the CrossValidation [[Measures\|measure]] (using the meanSquareError function, and a target of 0.1) with the LOLA sample selector to select new samples. It builds models for this output using the lssvmps model builder. The second output, named "outinverse", uses CrossValidation in combination with the MinMax measure, and uses the delaunay sample selector to select new samples for this output. It builds models using the model builder defined higher up in the xml hierarchy. As you can see, you can add multiple measures to one Output, so that each measure has something to say in deciding the accuracy of the models. However, you can only select one SampleSelector or AdaptiveModelBuilder.

	~~If an Output element does not contain any Measure elements, that output is modeled using the default configuration, which is the CrossValidation measure and target accuracy 0.001.~~		For information on how multiple measures are handled and the configuration options of each measure, see [[Measures]].

	For ~~more~~ information on how multiple measures are handled and the configuration options of each measure, see [[Measures]].

	=== Complex handling ===		=== Complex handling ===

Dgorissen: /* Default behavior, selected outputs modeled */

2009-09-29T18:08:16Z

Default behavior, selected outputs modeled

← Older revision		Revision as of 18:08, 29 September 2009
Line 7:		Line 7:
	== Default behavior, selected outputs modeled ==		== Default behavior, selected outputs modeled ==

	To change this default behavior, ~~one~~ must specify an Outputs tag in the Run configuration. If you want to ~~keep~~ the ~~default modeling behavior of the toolbox (CrossValidation~~ and ~~MinMax, 0~~.~~001 accuracy), but~~ you ~~don't~~ want to model ~~each output,~~ you ~~only have to add an Output subelement for each output that you want modeled.~~		To change this default behavior, you must specify an <Outputs> tag in the Run configuration. Inside you specify an <Output> tag for each output you want to model. As an example, the Academic2DTwice example has two outputs: 'out' and 'outinverse'. If you only want to model 'outinverse' you specify:

	~~An example of such a configuration for~~ the ~~Academic2DTwice test function is~~:		<source lang="xml">
			<Outputs>
			<Output name="outinverse" />
			</Outputs>
			</source>

			If on the other hand you want to model both you can delete the <Outputs> tag all together and fall back to the default behavior, or you can put:

	<source lang="xml">		<source lang="xml">

Dgorissen: /* Default behavior, all outputs modeled */

2009-09-29T18:03:46Z

Default behavior, all outputs modeled

← Older revision		Revision as of 18:03, 29 September 2009
Line 3:		Line 3:
	== Default behavior, all outputs modeled ==		== Default behavior, all outputs modeled ==

	If no Outputs tag is defined in the configuration file, all outputs are modeled and evaluated with the default CrossValidation ~~and MinMax (see~~ [[Measures]]~~) measure~~ and target accuracy 0.~~001~~. Complex outputs modeled directly.		If no Outputs tag is defined in the configuration file, all outputs are modeled and evaluated with the default CrossValidation [[Measures\|measure]] and target accuracy 0.01. Complex outputs are modeled directly.

	== Default behavior, selected outputs modeled ==		== Default behavior, selected outputs modeled ==

Admin: /* Default behavior, selected outputs modeled */

2009-09-19T16:11:13Z

Default behavior, selected outputs modeled

← Older revision		Revision as of 16:11, 19 September 2009
Line 11:		Line 11:
	An example of such a configuration for the Academic2DTwice test function is:		An example of such a configuration for the Academic2DTwice test function is:

			<source lang="xml">
	<~~pre><nowiki~~>
	<Outputs>		<Outputs>
	<Output name="out" />		<Output name="out" />
	<Output name="outinverse" />		<Output name="outinverse" />
	</Outputs>		</Outputs>
	</~~nowiki~~>~~</pre>~~		</source>


	== Custom behavior for each output ==		== Custom behavior for each output ==

Admin: /* Component customization */

2009-09-19T16:10:54Z

Component customization

← Older revision		Revision as of 16:10, 19 September 2009
Line 44:		Line 44:
	</Output>		</Output>
	</Outputs>		</Outputs>
	</~~nowiki></pre~~>		</source>

	This configuration models the first Output named "out" using the CrossValidation measure, but with increased target accuracy, and uses the gradient sample selector to select new samples. It builds models for this output using the svm model builder. The second output, named "outinverse", uses CrossValidation in combination with the MinMax measure, and uses the grid sample selector to select new samples for this output. It builds models using the model builder defined higher up in the xml hierarchy. As you can see, you can add multiple measures to one Output, so that each measure has something to say in deciding the accuracy of the models. However, you can only select one SampleSelector or AdaptiveModelBuilder.		This configuration models the first Output named "out" using the CrossValidation measure, but with increased target accuracy, and uses the gradient sample selector to select new samples. It builds models for this output using the svm model builder. The second output, named "outinverse", uses CrossValidation in combination with the MinMax measure, and uses the grid sample selector to select new samples for this output. It builds models using the model builder defined higher up in the xml hierarchy. As you can see, you can add multiple measures to one Output, so that each measure has something to say in deciding the accuracy of the models. However, you can only select one SampleSelector or AdaptiveModelBuilder.