TrainOGRLayersClassifier

Brief Description

Train a SVM classifier based on labeled geometries and a list of features to consider.

Long Description

This application trains a SVM classifier based on labeled geometries and a list of features to consider for classification. This application is deprecated, prefer using TrainVectorClassifier which offers access to all the classifiers.

Parameters

[param] -inshp <string> Name of the input shapefile. Mandatory: True. Default Value: ""
[param] -instats <string> XML file containing mean and variance of each feature.. Mandatory: True. Default Value: ""
[param] -outsvm <string> Output model filename.. Mandatory: True. Default Value: ""
[param] -cfield <string> Field containing the class id for supervision. Only geometries with this field available will be taken into account.. Mandatory: True. Default Value: "class"
[param] -inxml <string> Load otb application from xml file. Mandatory: False. Default Value: ""
[param] -outxml <string> Save otb application to xml file. Mandatory: False. Default Value: ""

[choice] -feat

Limitations

Experimental. For now only shapefiles are supported. Tuning of SVM classifier is not available.

Authors

David Youssefi during internship at CNES

Example of use

inshp: vectorData.shp
instats: meanVar.xml
outsvm: svmModel.svm
feat: perimeter
cfield: predicted