http://sumowiki.intec.ugent.be/index.php?title=Add_Model_Type&feed=atom&action=historyAdd Model Type - Revision history2024-03-29T07:21:01ZRevision history for this page on the wikiMediaWiki 1.35.4http://sumowiki.intec.ugent.be/index.php?title=Add_Model_Type&diff=5893&oldid=prevAdmin at 09:06, 17 March 20142014-03-17T09:06:06Z<p></p>
<table class="diff diff-contentalign-left diff-editfont-monospace" data-mw="interface">
<col class="diff-marker" />
<col class="diff-content" />
<col class="diff-marker" />
<col class="diff-content" />
<tr class="diff-title" lang="en">
<td colspan="2" style="background-color: #fff; color: #202122; text-align: center;">← Older revision</td>
<td colspan="2" style="background-color: #fff; color: #202122; text-align: center;">Revision as of 09:06, 17 March 2014</td>
</tr><tr><td colspan="2" class="diff-lineno" id="mw-diff-left-l14" >Line 14:</td>
<td colspan="2" class="diff-lineno">Line 14:</td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>To facilitate the setting of the model parameters (=hyperparameters) the SUMO Toolbox uses '''Model Builders'''. Finding the best model parameters given the data can be seen as a (constrained) optimization problem over the parameter space. Model Builders are implementations of '''optimization algorithms''' (pattern search, GA, simulated annealing, NSGA-II,...) which can be used for this task. For information on how to add a new Model Builder see [[Add Modeling Algorithm| Add a Model Builder]]. One single run of a model builder (e.g, doing one round of parameter optimization for SVM with pattern search) is referred to as one ''modeling iteration''.</div></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>To facilitate the setting of the model parameters (=hyperparameters) the SUMO Toolbox uses '''Model Builders'''. Finding the best model parameters given the data can be seen as a (constrained) optimization problem over the parameter space. Model Builders are implementations of '''optimization algorithms''' (pattern search, GA, simulated annealing, NSGA-II,...) which can be used for this task. For information on how to add a new Model Builder see [[Add Modeling Algorithm| Add a Model Builder]]. One single run of a model builder (e.g, doing one round of parameter optimization for SVM with pattern search) is referred to as one ''modeling iteration''.</div></td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td></tr>
<tr><td class='diff-marker'>−</td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;"><div>'''Model Factories''' provide an '''interface''' between the Model Builder and the Model. A Model Factory is responsible for reading the configuration options and telling the Model Builder what hyperparameters it has to optimize. In turn the Model Builder will instruct the Model Factory to construct an actual Model objects with the hyperparameters found during the optimization search. Without the Model Factory, each optimization algorithm would have to be manually customized to work with a particular model. This is reflected in the configuration XML: each <del class="diffchange diffchange-inline">AdaptiveModelBuilder </del>configuration section contains a Model Factory as well as the ModelBuilder itself. For example the 'anngenetic' <del class="diffchange diffchange-inline">AdaptiveModelBuilder</del>, is a combination of a 'GeneticModelBuilder' and an 'ANNFactory'.</div></td><td class='diff-marker'>+</td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;"><div>'''Model Factories''' provide an '''interface''' between the Model Builder and the Model. A Model Factory is responsible for reading the configuration options and telling the Model Builder what hyperparameters it has to optimize. In turn the Model Builder will instruct the Model Factory to construct an actual Model objects with the hyperparameters found during the optimization search. Without the Model Factory, each optimization algorithm would have to be manually customized to work with a particular model. This is reflected in the configuration XML: each <ins class="diffchange diffchange-inline">ModelBuilder </ins>configuration section contains a Model Factory as well as the ModelBuilder itself. For example the 'anngenetic' <ins class="diffchange diffchange-inline">ModelBuilder</ins>, is a combination of a 'GeneticModelBuilder' and an 'ANNFactory'.</div></td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>The simplest way to see how this work is by looking at some of the Model and Model Factories included in the SUMO Toolbox such as SplineFactory and SplineModel.</div></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>The simplest way to see how this work is by looking at some of the Model and Model Factories included in the SUMO Toolbox such as SplineFactory and SplineModel.</div></td></tr>
</table>Adminhttp://sumowiki.intec.ugent.be/index.php?title=Add_Model_Type&diff=5647&oldid=prevMnguyen: /* Models, Model builders, and Factories */2012-01-19T17:24:39Z<p><span dir="auto"><span class="autocomment">Models, Model builders, and Factories</span></span></p>
<table class="diff diff-contentalign-left diff-editfont-monospace" data-mw="interface">
<col class="diff-marker" />
<col class="diff-content" />
<col class="diff-marker" />
<col class="diff-content" />
<tr class="diff-title" lang="en">
<td colspan="2" style="background-color: #fff; color: #202122; text-align: center;">← Older revision</td>
<td colspan="2" style="background-color: #fff; color: #202122; text-align: center;">Revision as of 17:24, 19 January 2012</td>
</tr><tr><td colspan="2" class="diff-lineno" id="mw-diff-left-l14" >Line 14:</td>
<td colspan="2" class="diff-lineno">Line 14:</td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>To facilitate the setting of the model parameters (=hyperparameters) the SUMO Toolbox uses '''Model Builders'''. Finding the best model parameters given the data can be seen as a (constrained) optimization problem over the parameter space. Model Builders are implementations of '''optimization algorithms''' (pattern search, GA, simulated annealing, NSGA-II,...) which can be used for this task. For information on how to add a new Model Builder see [[Add Modeling Algorithm| Add a Model Builder]]. One single run of a model builder (e.g, doing one round of parameter optimization for SVM with pattern search) is referred to as one ''modeling iteration''.</div></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>To facilitate the setting of the model parameters (=hyperparameters) the SUMO Toolbox uses '''Model Builders'''. Finding the best model parameters given the data can be seen as a (constrained) optimization problem over the parameter space. Model Builders are implementations of '''optimization algorithms''' (pattern search, GA, simulated annealing, NSGA-II,...) which can be used for this task. For information on how to add a new Model Builder see [[Add Modeling Algorithm| Add a Model Builder]]. One single run of a model builder (e.g, doing one round of parameter optimization for SVM with pattern search) is referred to as one ''modeling iteration''.</div></td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td></tr>
<tr><td class='diff-marker'>−</td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;"><div>'''Model Factories''' provide an '''interface''' between the Model Builder and the Model. A Model Factory is responsible for reading the configuration options and telling the Model Builder what hyperparameters it has to optimize. In turn the Model Builder will instruct the Model Factory to construct an actual Model objects with the hyperparameters found during the optimization search. Without the Model Factory, each optimization algorithm would have to be manually customized to work with a particular model. This is reflected in the configuration XML<del class="diffchange diffchange-inline">, </del>each AdaptiveModelBuilder configuration section contains a Model Factory as well as the ModelBuilder itself. For example the 'anngenetic' AdaptiveModelBuilder, is a combination of a 'GeneticModelBuilder' and an 'ANNFactory'.</div></td><td class='diff-marker'>+</td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;"><div>'''Model Factories''' provide an '''interface''' between the Model Builder and the Model. A Model Factory is responsible for reading the configuration options and telling the Model Builder what hyperparameters it has to optimize. In turn the Model Builder will instruct the Model Factory to construct an actual Model objects with the hyperparameters found during the optimization search. Without the Model Factory, each optimization algorithm would have to be manually customized to work with a particular model. This is reflected in the configuration XML<ins class="diffchange diffchange-inline">: </ins>each AdaptiveModelBuilder configuration section contains a Model Factory as well as the ModelBuilder itself. For example the 'anngenetic' AdaptiveModelBuilder, is a combination of a 'GeneticModelBuilder' and an 'ANNFactory'.</div></td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>The simplest way to see how this work is by looking at some of the Model and Model Factories included in the SUMO Toolbox such as SplineFactory and SplineModel.</div></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>The simplest way to see how this work is by looking at some of the Model and Model Factories included in the SUMO Toolbox such as SplineFactory and SplineModel.</div></td></tr>
</table>Mnguyenhttp://sumowiki.intec.ugent.be/index.php?title=Add_Model_Type&diff=5646&oldid=prevMnguyen: /* Models, Model builders, and Factories */2012-01-19T17:24:22Z<p><span dir="auto"><span class="autocomment">Models, Model builders, and Factories</span></span></p>
<table class="diff diff-contentalign-left diff-editfont-monospace" data-mw="interface">
<col class="diff-marker" />
<col class="diff-content" />
<col class="diff-marker" />
<col class="diff-content" />
<tr class="diff-title" lang="en">
<td colspan="2" style="background-color: #fff; color: #202122; text-align: center;">← Older revision</td>
<td colspan="2" style="background-color: #fff; color: #202122; text-align: center;">Revision as of 17:24, 19 January 2012</td>
</tr><tr><td colspan="2" class="diff-lineno" id="mw-diff-left-l14" >Line 14:</td>
<td colspan="2" class="diff-lineno">Line 14:</td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>To facilitate the setting of the model parameters (=hyperparameters) the SUMO Toolbox uses '''Model Builders'''. Finding the best model parameters given the data can be seen as a (constrained) optimization problem over the parameter space. Model Builders are implementations of '''optimization algorithms''' (pattern search, GA, simulated annealing, NSGA-II,...) which can be used for this task. For information on how to add a new Model Builder see [[Add Modeling Algorithm| Add a Model Builder]]. One single run of a model builder (e.g, doing one round of parameter optimization for SVM with pattern search) is referred to as one ''modeling iteration''.</div></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>To facilitate the setting of the model parameters (=hyperparameters) the SUMO Toolbox uses '''Model Builders'''. Finding the best model parameters given the data can be seen as a (constrained) optimization problem over the parameter space. Model Builders are implementations of '''optimization algorithms''' (pattern search, GA, simulated annealing, NSGA-II,...) which can be used for this task. For information on how to add a new Model Builder see [[Add Modeling Algorithm| Add a Model Builder]]. One single run of a model builder (e.g, doing one round of parameter optimization for SVM with pattern search) is referred to as one ''modeling iteration''.</div></td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td></tr>
<tr><td class='diff-marker'>−</td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;"><div>'''Model Factories''' provide an '''interface''' between the Model Builder and the Model. A Model Factory is responsible for reading the configuration options and telling the Model Builder what hyperparameters it has to optimize. In turn the Model Builder will instruct the Model Factory to construct an actual Model objects with the hyperparameters found during the optimization search. Without the Model Factory, each optimization algorithm would have to be manually customized to work with a particular model. <del class="diffchange diffchange-inline">Therefore, </del>in the configuration XML, each AdaptiveModelBuilder contains a Model Factory <del class="diffchange diffchange-inline">(</del>as <del class="diffchange diffchange-inline">you can see from </del>the <del class="diffchange diffchange-inline">XML configuration)</del>. For example the anngenetic AdaptiveModelBuilder, is a combination of a GeneticModelBuilder and an ANNFactory.</div></td><td class='diff-marker'>+</td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;"><div>'''Model Factories''' provide an '''interface''' between the Model Builder and the Model. A Model Factory is responsible for reading the configuration options and telling the Model Builder what hyperparameters it has to optimize. In turn the Model Builder will instruct the Model Factory to construct an actual Model objects with the hyperparameters found during the optimization search. Without the Model Factory, each optimization algorithm would have to be manually customized to work with a particular model. <ins class="diffchange diffchange-inline">This is reflected </ins>in the configuration XML, each AdaptiveModelBuilder <ins class="diffchange diffchange-inline">configuration section </ins>contains a Model Factory as <ins class="diffchange diffchange-inline">well as </ins>the <ins class="diffchange diffchange-inline">ModelBuilder itself</ins>. For example the <ins class="diffchange diffchange-inline">'</ins>anngenetic<ins class="diffchange diffchange-inline">' </ins>AdaptiveModelBuilder, is a combination of a <ins class="diffchange diffchange-inline">'</ins>GeneticModelBuilder<ins class="diffchange diffchange-inline">' </ins>and an <ins class="diffchange diffchange-inline">'</ins>ANNFactory<ins class="diffchange diffchange-inline">'</ins>.</div></td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>The simplest way to see how this work is by looking at some of the Model and Model Factories included in the SUMO Toolbox such as SplineFactory and SplineModel.</div></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>The simplest way to see how this work is by looking at some of the Model and Model Factories included in the SUMO Toolbox such as SplineFactory and SplineModel.</div></td></tr>
</table>Mnguyenhttp://sumowiki.intec.ugent.be/index.php?title=Add_Model_Type&diff=5645&oldid=prevMnguyen: /* Models, Model builders, and Factories */2012-01-19T17:22:35Z<p><span dir="auto"><span class="autocomment">Models, Model builders, and Factories</span></span></p>
<table class="diff diff-contentalign-left diff-editfont-monospace" data-mw="interface">
<col class="diff-marker" />
<col class="diff-content" />
<col class="diff-marker" />
<col class="diff-content" />
<tr class="diff-title" lang="en">
<td colspan="2" style="background-color: #fff; color: #202122; text-align: center;">← Older revision</td>
<td colspan="2" style="background-color: #fff; color: #202122; text-align: center;">Revision as of 17:22, 19 January 2012</td>
</tr><tr><td colspan="2" class="diff-lineno" id="mw-diff-left-l14" >Line 14:</td>
<td colspan="2" class="diff-lineno">Line 14:</td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>To facilitate the setting of the model parameters (=hyperparameters) the SUMO Toolbox uses '''Model Builders'''. Finding the best model parameters given the data can be seen as a (constrained) optimization problem over the parameter space. Model Builders are implementations of '''optimization algorithms''' (pattern search, GA, simulated annealing, NSGA-II,...) which can be used for this task. For information on how to add a new Model Builder see [[Add Modeling Algorithm| Add a Model Builder]]. One single run of a model builder (e.g, doing one round of parameter optimization for SVM with pattern search) is referred to as one ''modeling iteration''.</div></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>To facilitate the setting of the model parameters (=hyperparameters) the SUMO Toolbox uses '''Model Builders'''. Finding the best model parameters given the data can be seen as a (constrained) optimization problem over the parameter space. Model Builders are implementations of '''optimization algorithms''' (pattern search, GA, simulated annealing, NSGA-II,...) which can be used for this task. For information on how to add a new Model Builder see [[Add Modeling Algorithm| Add a Model Builder]]. One single run of a model builder (e.g, doing one round of parameter optimization for SVM with pattern search) is referred to as one ''modeling iteration''.</div></td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td></tr>
<tr><td class='diff-marker'>−</td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;"><div>'''Model Factories''' provide an '''interface''' between the Model Builder and the Model. A Model Factory is responsible for reading the configuration options and telling the Model Builder what hyperparameters it has to optimize. In turn the Model Builder will instruct the Model Factory to construct an actual Model objects with the <del class="diffchange diffchange-inline">optimzed </del>hyperparameters found during the optimization search. Without the Model Factory, each optimization algorithm would have to be manually customized to work with a particular model. Therefore, in the configuration XML, each AdaptiveModelBuilder contains a Model Factory (as you can see from the XML configuration). For example the anngenetic AdaptiveModelBuilder, is a combination of a GeneticModelBuilder and an ANNFactory.</div></td><td class='diff-marker'>+</td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;"><div>'''Model Factories''' provide an '''interface''' between the Model Builder and the Model. A Model Factory is responsible for reading the configuration options and telling the Model Builder what hyperparameters it has to optimize. In turn the Model Builder will instruct the Model Factory to construct an actual Model objects with the hyperparameters found during the optimization search. Without the Model Factory, each optimization algorithm would have to be manually customized to work with a particular model. Therefore, in the configuration XML, each AdaptiveModelBuilder contains a Model Factory (as you can see from the XML configuration). For example the anngenetic AdaptiveModelBuilder, is a combination of a GeneticModelBuilder and an ANNFactory.</div></td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>The simplest way to see how this work is by looking at some of the Model and Model Factories included in the SUMO Toolbox such as SplineFactory and SplineModel.</div></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>The simplest way to see how this work is by looking at some of the Model and Model Factories included in the SUMO Toolbox such as SplineFactory and SplineModel.</div></td></tr>
</table>Mnguyenhttp://sumowiki.intec.ugent.be/index.php?title=Add_Model_Type&diff=5644&oldid=prevMnguyen: /* Models, Model builders, and Factories */2012-01-19T17:21:38Z<p><span dir="auto"><span class="autocomment">Models, Model builders, and Factories</span></span></p>
<table class="diff diff-contentalign-left diff-editfont-monospace" data-mw="interface">
<col class="diff-marker" />
<col class="diff-content" />
<col class="diff-marker" />
<col class="diff-content" />
<tr class="diff-title" lang="en">
<td colspan="2" style="background-color: #fff; color: #202122; text-align: center;">← Older revision</td>
<td colspan="2" style="background-color: #fff; color: #202122; text-align: center;">Revision as of 17:21, 19 January 2012</td>
</tr><tr><td colspan="2" class="diff-lineno" id="mw-diff-left-l14" >Line 14:</td>
<td colspan="2" class="diff-lineno">Line 14:</td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>To facilitate the setting of the model parameters (=hyperparameters) the SUMO Toolbox uses '''Model Builders'''. Finding the best model parameters given the data can be seen as a (constrained) optimization problem over the parameter space. Model Builders are implementations of '''optimization algorithms''' (pattern search, GA, simulated annealing, NSGA-II,...) which can be used for this task. For information on how to add a new Model Builder see [[Add Modeling Algorithm| Add a Model Builder]]. One single run of a model builder (e.g, doing one round of parameter optimization for SVM with pattern search) is referred to as one ''modeling iteration''.</div></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>To facilitate the setting of the model parameters (=hyperparameters) the SUMO Toolbox uses '''Model Builders'''. Finding the best model parameters given the data can be seen as a (constrained) optimization problem over the parameter space. Model Builders are implementations of '''optimization algorithms''' (pattern search, GA, simulated annealing, NSGA-II,...) which can be used for this task. For information on how to add a new Model Builder see [[Add Modeling Algorithm| Add a Model Builder]]. One single run of a model builder (e.g, doing one round of parameter optimization for SVM with pattern search) is referred to as one ''modeling iteration''.</div></td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td></tr>
<tr><td class='diff-marker'>−</td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;"><div>'''Model Factories''' provide an '''interface''' between the Model Builder and the Model. A Model Factory is responsible for reading the configuration options and telling the Model Builder what hyperparameters it has to optimize. In turn the Model Builder will instruct the Model Factory to construct actual Model objects with the hyperparameters <del class="diffchange diffchange-inline">it has </del>found. Without the Model Factory, each optimization algorithm would have to be manually customized to work with a particular model. Therefore, in the configuration XML, each AdaptiveModelBuilder contains a Model Factory (as you can see from the XML configuration). For example the anngenetic AdaptiveModelBuilder, is a combination of a GeneticModelBuilder and an ANNFactory.</div></td><td class='diff-marker'>+</td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;"><div>'''Model Factories''' provide an '''interface''' between the Model Builder and the Model. A Model Factory is responsible for reading the configuration options and telling the Model Builder what hyperparameters it has to optimize. In turn the Model Builder will instruct the Model Factory to construct <ins class="diffchange diffchange-inline">an </ins>actual Model objects with the <ins class="diffchange diffchange-inline">optimzed </ins>hyperparameters found <ins class="diffchange diffchange-inline">during the optimization search</ins>. Without the Model Factory, each optimization algorithm would have to be manually customized to work with a particular model. Therefore, in the configuration XML, each AdaptiveModelBuilder contains a Model Factory (as you can see from the XML configuration). For example the anngenetic AdaptiveModelBuilder, is a combination of a GeneticModelBuilder and an ANNFactory.</div></td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>The simplest way to see how this work is by looking at some of the Model and Model Factories included in the SUMO Toolbox such as SplineFactory and SplineModel.</div></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>The simplest way to see how this work is by looking at some of the Model and Model Factories included in the SUMO Toolbox such as SplineFactory and SplineModel.</div></td></tr>
</table>Mnguyenhttp://sumowiki.intec.ugent.be/index.php?title=Add_Model_Type&diff=5643&oldid=prevMnguyen: /* Models, Model builders, and Factories */2012-01-19T17:20:17Z<p><span dir="auto"><span class="autocomment">Models, Model builders, and Factories</span></span></p>
<table class="diff diff-contentalign-left diff-editfont-monospace" data-mw="interface">
<col class="diff-marker" />
<col class="diff-content" />
<col class="diff-marker" />
<col class="diff-content" />
<tr class="diff-title" lang="en">
<td colspan="2" style="background-color: #fff; color: #202122; text-align: center;">← Older revision</td>
<td colspan="2" style="background-color: #fff; color: #202122; text-align: center;">Revision as of 17:20, 19 January 2012</td>
</tr><tr><td colspan="2" class="diff-lineno" id="mw-diff-left-l14" >Line 14:</td>
<td colspan="2" class="diff-lineno">Line 14:</td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>To facilitate the setting of the model parameters (=hyperparameters) the SUMO Toolbox uses '''Model Builders'''. Finding the best model parameters given the data can be seen as a (constrained) optimization problem over the parameter space. Model Builders are implementations of '''optimization algorithms''' (pattern search, GA, simulated annealing, NSGA-II,...) which can be used for this task. For information on how to add a new Model Builder see [[Add Modeling Algorithm| Add a Model Builder]]. One single run of a model builder (e.g, doing one round of parameter optimization for SVM with pattern search) is referred to as one ''modeling iteration''.</div></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>To facilitate the setting of the model parameters (=hyperparameters) the SUMO Toolbox uses '''Model Builders'''. Finding the best model parameters given the data can be seen as a (constrained) optimization problem over the parameter space. Model Builders are implementations of '''optimization algorithms''' (pattern search, GA, simulated annealing, NSGA-II,...) which can be used for this task. For information on how to add a new Model Builder see [[Add Modeling Algorithm| Add a Model Builder]]. One single run of a model builder (e.g, doing one round of parameter optimization for SVM with pattern search) is referred to as one ''modeling iteration''.</div></td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td></tr>
<tr><td class='diff-marker'>−</td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;"><div>'''Model Factories''' provide an '''interface''' between the Model Builder and the Model. A Model Factory is responsible for reading the configuration options and telling the Model Builder what hyperparameters it has to optimize. <del class="diffchange diffchange-inline">And in </del>turn the Model Builder will instruct the Model Factory to construct actual Model objects with the hyperparameters it has found. Without the Model Factory, each optimization algorithm would have to be manually customized to work with a particular model. Therefore, in the configuration XML, each AdaptiveModelBuilder contains a Model Factory (as you can see from the XML configuration). For example the anngenetic AdaptiveModelBuilder, is a combination of a GeneticModelBuilder and an ANNFactory.</div></td><td class='diff-marker'>+</td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;"><div>'''Model Factories''' provide an '''interface''' between the Model Builder and the Model. A Model Factory is responsible for reading the configuration options and telling the Model Builder what hyperparameters it has to optimize. <ins class="diffchange diffchange-inline">In </ins>turn the Model Builder will instruct the Model Factory to construct actual Model objects with the hyperparameters it has found. Without the Model Factory, each optimization algorithm would have to be manually customized to work with a particular model. Therefore, in the configuration XML, each AdaptiveModelBuilder contains a Model Factory (as you can see from the XML configuration). For example the anngenetic AdaptiveModelBuilder, is a combination of a GeneticModelBuilder and an ANNFactory.</div></td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>The simplest way to see how this work is by looking at some of the Model and Model Factories included in the SUMO Toolbox such as SplineFactory and SplineModel.</div></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>The simplest way to see how this work is by looking at some of the Model and Model Factories included in the SUMO Toolbox such as SplineFactory and SplineModel.</div></td></tr>
</table>Mnguyenhttp://sumowiki.intec.ugent.be/index.php?title=Add_Model_Type&diff=5642&oldid=prevMnguyen: /* Models, Model builders, and Factories */2012-01-19T17:19:55Z<p><span dir="auto"><span class="autocomment">Models, Model builders, and Factories</span></span></p>
<table class="diff diff-contentalign-left diff-editfont-monospace" data-mw="interface">
<col class="diff-marker" />
<col class="diff-content" />
<col class="diff-marker" />
<col class="diff-content" />
<tr class="diff-title" lang="en">
<td colspan="2" style="background-color: #fff; color: #202122; text-align: center;">← Older revision</td>
<td colspan="2" style="background-color: #fff; color: #202122; text-align: center;">Revision as of 17:19, 19 January 2012</td>
</tr><tr><td colspan="2" class="diff-lineno" id="mw-diff-left-l14" >Line 14:</td>
<td colspan="2" class="diff-lineno">Line 14:</td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>To facilitate the setting of the model parameters (=hyperparameters) the SUMO Toolbox uses '''Model Builders'''. Finding the best model parameters given the data can be seen as a (constrained) optimization problem over the parameter space. Model Builders are implementations of '''optimization algorithms''' (pattern search, GA, simulated annealing, NSGA-II,...) which can be used for this task. For information on how to add a new Model Builder see [[Add Modeling Algorithm| Add a Model Builder]]. One single run of a model builder (e.g, doing one round of parameter optimization for SVM with pattern search) is referred to as one ''modeling iteration''.</div></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>To facilitate the setting of the model parameters (=hyperparameters) the SUMO Toolbox uses '''Model Builders'''. Finding the best model parameters given the data can be seen as a (constrained) optimization problem over the parameter space. Model Builders are implementations of '''optimization algorithms''' (pattern search, GA, simulated annealing, NSGA-II,...) which can be used for this task. For information on how to add a new Model Builder see [[Add Modeling Algorithm| Add a Model Builder]]. One single run of a model builder (e.g, doing one round of parameter optimization for SVM with pattern search) is referred to as one ''modeling iteration''.</div></td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td></tr>
<tr><td class='diff-marker'>−</td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;"><div>'''Model Factories''' provide an '''interface''' between the Model Builder and the Model. A Model Factory is responsible for reading the configuration options and <del class="diffchange diffchange-inline">tells </del>the Model Builder what hyperparameters it has to optimize. And in turn the Model Builder will instruct the Model Factory to construct actual Model objects with the hyperparameters it has found. Without the Model Factory, each optimization algorithm would have to be manually customized to work with a particular model. Therefore, in the configuration XML, each AdaptiveModelBuilder contains a Model Factory (as you can see from the XML configuration). For example the anngenetic AdaptiveModelBuilder, is a combination of a GeneticModelBuilder and an ANNFactory.</div></td><td class='diff-marker'>+</td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;"><div>'''Model Factories''' provide an '''interface''' between the Model Builder and the Model. A Model Factory is responsible for reading the configuration options and <ins class="diffchange diffchange-inline">telling </ins>the Model Builder what hyperparameters it has to optimize. And in turn the Model Builder will instruct the Model Factory to construct actual Model objects with the hyperparameters it has found. Without the Model Factory, each optimization algorithm would have to be manually customized to work with a particular model. Therefore, in the configuration XML, each AdaptiveModelBuilder contains a Model Factory (as you can see from the XML configuration). For example the anngenetic AdaptiveModelBuilder, is a combination of a GeneticModelBuilder and an ANNFactory.</div></td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>The simplest way to see how this work is by looking at some of the Model and Model Factories included in the SUMO Toolbox such as SplineFactory and SplineModel.</div></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>The simplest way to see how this work is by looking at some of the Model and Model Factories included in the SUMO Toolbox such as SplineFactory and SplineModel.</div></td></tr>
</table>Mnguyenhttp://sumowiki.intec.ugent.be/index.php?title=Add_Model_Type&diff=5641&oldid=prevMnguyen: /* Models, Model builders, and Factories */2012-01-19T17:19:17Z<p><span dir="auto"><span class="autocomment">Models, Model builders, and Factories</span></span></p>
<table class="diff diff-contentalign-left diff-editfont-monospace" data-mw="interface">
<col class="diff-marker" />
<col class="diff-content" />
<col class="diff-marker" />
<col class="diff-content" />
<tr class="diff-title" lang="en">
<td colspan="2" style="background-color: #fff; color: #202122; text-align: center;">← Older revision</td>
<td colspan="2" style="background-color: #fff; color: #202122; text-align: center;">Revision as of 17:19, 19 January 2012</td>
</tr><tr><td colspan="2" class="diff-lineno" id="mw-diff-left-l12" >Line 12:</td>
<td colspan="2" class="diff-lineno">Line 12:</td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>'''A Model''' is an implementation of a '''machine learning algorithm''', such Artificial Neural Networks (ANN) or Support Vector Machines (SVM). When a Model object is constructed when running the SUMO toolbox, its hyperparameters have to be set. This is where the difficulty usually lies. Setting good values for the model parameters to accurately model the data requires finding the right balance between the bias/variance tradeoff and in many cases this is more of an art than a science.</div></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>'''A Model''' is an implementation of a '''machine learning algorithm''', such Artificial Neural Networks (ANN) or Support Vector Machines (SVM). When a Model object is constructed when running the SUMO toolbox, its hyperparameters have to be set. This is where the difficulty usually lies. Setting good values for the model parameters to accurately model the data requires finding the right balance between the bias/variance tradeoff and in many cases this is more of an art than a science.</div></td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td></tr>
<tr><td class='diff-marker'>−</td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;"><div>To facilitate the setting of the model parameters (=hyperparameters) the SUMO Toolbox uses '''Model Builders'''. Finding the best model parameters given the data can be seen as a (constrained) optimization problem over the parameter space <del class="diffchange diffchange-inline">and </del>Model Builders are implementations of '''optimization algorithms''' (pattern search, GA, simulated annealing, NSGA-II,...) which can be used for this task. For information on how to add a new Model Builder see [[Add Modeling Algorithm| Add a Model Builder]]. One single run of a model builder (e.g, doing one round of parameter optimization for SVM with pattern search) is referred to as one ''modeling iteration''.</div></td><td class='diff-marker'>+</td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;"><div>To facilitate the setting of the model parameters (=hyperparameters) the SUMO Toolbox uses '''Model Builders'''. Finding the best model parameters given the data can be seen as a (constrained) optimization problem over the parameter space<ins class="diffchange diffchange-inline">. </ins>Model Builders are implementations of '''optimization algorithms''' (pattern search, GA, simulated annealing, NSGA-II,...) which can be used for this task. For information on how to add a new Model Builder see [[Add Modeling Algorithm| Add a Model Builder]]. One single run of a model builder (e.g, doing one round of parameter optimization for SVM with pattern search) is referred to as one ''modeling iteration''.</div></td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>'''Model Factories''' provide an '''interface''' between the Model Builder and the Model. A Model Factory is responsible for reading the configuration options and tells the Model Builder what hyperparameters it has to optimize. And in turn the Model Builder will instruct the Model Factory to construct actual Model objects with the hyperparameters it has found. Without the Model Factory, each optimization algorithm would have to be manually customized to work with a particular model. Therefore, in the configuration XML, each AdaptiveModelBuilder contains a Model Factory (as you can see from the XML configuration). For example the anngenetic AdaptiveModelBuilder, is a combination of a GeneticModelBuilder and an ANNFactory.</div></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>'''Model Factories''' provide an '''interface''' between the Model Builder and the Model. A Model Factory is responsible for reading the configuration options and tells the Model Builder what hyperparameters it has to optimize. And in turn the Model Builder will instruct the Model Factory to construct actual Model objects with the hyperparameters it has found. Without the Model Factory, each optimization algorithm would have to be manually customized to work with a particular model. Therefore, in the configuration XML, each AdaptiveModelBuilder contains a Model Factory (as you can see from the XML configuration). For example the anngenetic AdaptiveModelBuilder, is a combination of a GeneticModelBuilder and an ANNFactory.</div></td></tr>
</table>Mnguyenhttp://sumowiki.intec.ugent.be/index.php?title=Add_Model_Type&diff=5640&oldid=prevMnguyen: /* Models, Model builders, and Factories */2012-01-19T17:18:39Z<p><span dir="auto"><span class="autocomment">Models, Model builders, and Factories</span></span></p>
<table class="diff diff-contentalign-left diff-editfont-monospace" data-mw="interface">
<col class="diff-marker" />
<col class="diff-content" />
<col class="diff-marker" />
<col class="diff-content" />
<tr class="diff-title" lang="en">
<td colspan="2" style="background-color: #fff; color: #202122; text-align: center;">← Older revision</td>
<td colspan="2" style="background-color: #fff; color: #202122; text-align: center;">Revision as of 17:18, 19 January 2012</td>
</tr><tr><td colspan="2" class="diff-lineno" id="mw-diff-left-l10" >Line 10:</td>
<td colspan="2" class="diff-lineno">Line 10:</td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>It is very important to first grasp the difference between Model builders, Model Factories and Models. Remember that each model type as a set of parameters that control the complexity of the model. For example a polynomial model has a degree parameter, an SVM has a kernel function, Kriging has theta parameters, etc. We refer to these parameters as hyperparameters or model parameters. In order to generate a good model you need to search for a good set of model parameters (with respect to a [[Measures|Measure]]). In essence this is an optimization problem in model parameter space (= hyperparameter space).</div></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>It is very important to first grasp the difference between Model builders, Model Factories and Models. Remember that each model type as a set of parameters that control the complexity of the model. For example a polynomial model has a degree parameter, an SVM has a kernel function, Kriging has theta parameters, etc. We refer to these parameters as hyperparameters or model parameters. In order to generate a good model you need to search for a good set of model parameters (with respect to a [[Measures|Measure]]). In essence this is an optimization problem in model parameter space (= hyperparameter space).</div></td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td></tr>
<tr><td class='diff-marker'>−</td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;"><div>'''A Model''' is an implementation of a '''machine learning algorithm''', such Artificial Neural Networks (ANN) or Support Vector Machines (SVM). When a Model object is constructed when running the SUMO toolbox, its hyperparameters have to be set. This is where the difficulty usually lies. Setting good values for the model parameters requires finding the right balance between the bias/variance tradeoff and in many cases this is more of an art than a science.</div></td><td class='diff-marker'>+</td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;"><div>'''A Model''' is an implementation of a '''machine learning algorithm''', such Artificial Neural Networks (ANN) or Support Vector Machines (SVM). When a Model object is constructed when running the SUMO toolbox, its hyperparameters have to be set. This is where the difficulty usually lies. Setting good values for the model parameters <ins class="diffchange diffchange-inline">to accurately model the data </ins>requires finding the right balance between the bias/variance tradeoff and in many cases this is more of an art than a science.</div></td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>To facilitate the setting of the model parameters (=hyperparameters) the SUMO Toolbox uses '''Model Builders'''. Finding the best model parameters given the data can be seen as a (constrained) optimization problem over the parameter space and Model Builders are implementations of '''optimization algorithms''' (pattern search, GA, simulated annealing, NSGA-II,...) which can be used for this task. For information on how to add a new Model Builder see [[Add Modeling Algorithm| Add a Model Builder]]. One single run of a model builder (e.g, doing one round of parameter optimization for SVM with pattern search) is referred to as one ''modeling iteration''.</div></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>To facilitate the setting of the model parameters (=hyperparameters) the SUMO Toolbox uses '''Model Builders'''. Finding the best model parameters given the data can be seen as a (constrained) optimization problem over the parameter space and Model Builders are implementations of '''optimization algorithms''' (pattern search, GA, simulated annealing, NSGA-II,...) which can be used for this task. For information on how to add a new Model Builder see [[Add Modeling Algorithm| Add a Model Builder]]. One single run of a model builder (e.g, doing one round of parameter optimization for SVM with pattern search) is referred to as one ''modeling iteration''.</div></td></tr>
</table>Mnguyenhttp://sumowiki.intec.ugent.be/index.php?title=Add_Model_Type&diff=5639&oldid=prevMnguyen: /* Models, Model builders, and Factories */2012-01-19T17:17:08Z<p><span dir="auto"><span class="autocomment">Models, Model builders, and Factories</span></span></p>
<table class="diff diff-contentalign-left diff-editfont-monospace" data-mw="interface">
<col class="diff-marker" />
<col class="diff-content" />
<col class="diff-marker" />
<col class="diff-content" />
<tr class="diff-title" lang="en">
<td colspan="2" style="background-color: #fff; color: #202122; text-align: center;">← Older revision</td>
<td colspan="2" style="background-color: #fff; color: #202122; text-align: center;">Revision as of 17:17, 19 January 2012</td>
</tr><tr><td colspan="2" class="diff-lineno" id="mw-diff-left-l5" >Line 5:</td>
<td colspan="2" class="diff-lineno">Line 5:</td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>== Models, Model builders, and Factories ==</div></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>== Models, Model builders, and Factories ==</div></td></tr>
<tr><td colspan="2"> </td><td class='diff-marker'>+</td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;"><div><ins style="font-weight: bold; text-decoration: none;"></ins></div></td></tr>
<tr><td colspan="2"> </td><td class='diff-marker'>+</td><td style="color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;"><div><ins style="font-weight: bold; text-decoration: none;">[[Image:Modelfactory.png|thumb|Schematic representation of the relationship between ModelBuilder, ModelFactory, Model. Click on the picture for a larger version.]]</ins></div></td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"></td></tr>
<tr><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>It is very important to first grasp the difference between Model builders, Model Factories and Models. Remember that each model type as a set of parameters that control the complexity of the model. For example a polynomial model has a degree parameter, an SVM has a kernel function, Kriging has theta parameters, etc. We refer to these parameters as hyperparameters or model parameters. In order to generate a good model you need to search for a good set of model parameters (with respect to a [[Measures|Measure]]). In essence this is an optimization problem in model parameter space (= hyperparameter space).</div></td><td class='diff-marker'> </td><td style="background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;"><div>It is very important to first grasp the difference between Model builders, Model Factories and Models. Remember that each model type as a set of parameters that control the complexity of the model. For example a polynomial model has a degree parameter, an SVM has a kernel function, Kriging has theta parameters, etc. We refer to these parameters as hyperparameters or model parameters. In order to generate a good model you need to search for a good set of model parameters (with respect to a [[Measures|Measure]]). In essence this is an optimization problem in model parameter space (= hyperparameter space).</div></td></tr>
</table>Mnguyen