TrainVectorClassifier¶
Train a classifier based on labeled geometries and a list of features to consider.
Description¶
This application trains a classifier based on labeled geometries and a list of features to consider for classification. This application is based on LibSVM, OpenCV Machine Learning (2.3.1 and later), and Shark ML The output of this application is a text model file, whose format corresponds to the ML model type chosen. There are no image or vector data outputs created.
Parameters¶
Input and output data¶
This group of parameters allows setting input and output data.
Input Vector Data -io.vd vectorfile1 vectorfile2... Mandatory
Input geometries used for training (note: all geometries from the layer will be used)
Input XML image statistics file -io.stats filename [dtype]
XML file containing mean and variance of each feature.
Output model -io.out filename [dtype] Mandatory
Output file containing the model estimated (.txt format).
Output confusion matrix or contingency table -io.confmatout filename [dtype]
Output file containing the confusion matrix or contingency table (.csv format).The contingency table is output when we unsupervised algorithms is used otherwise the confusion matrix is output.
Layer Index -layer int Default value: 0
Index of the layer to use in the input vector file.
Field names for training features -feat string1 string2...
List of field names in the input vector data to be used as features for training.
Validation data¶
This group of parameters defines validation data.
Validation Vector Data -valid.vd vectorfile1 vectorfile2...
Geometries used for validation (must contain the same fields used for training, all geometries from the layer will be used)
Layer Index -valid.layer int Default value: 0
Index of the layer to use in the validation vector file.
Field containing the class integer label for supervision -cfield string
Field containing the class id for supervision. The values in this field shall be cast into integers. Only geometries with this field available will be taken into account.
Verbose mode -v bool Default value: true
Verbose mode, display the contingency table result.
Classifier to use for the training -classifier [libsvm|boost|dt|ann|bayes|rf|knn|sharkrf|sharkkm] Default value: libsvm
Choice of the classifier to use for the training.
LibSVM classifier
This group of parameters allows setting SVM classifier parameters.Boost classifier
http://docs.opencv.org/modules/ml/doc/boosting.htmlDecision Tree classifier
http://docs.opencv.org/modules/ml/doc/decision_trees.htmlArtificial Neural Network classifier
http://docs.opencv.org/modules/ml/doc/neural_networks.htmlNormal Bayes classifier
http://docs.opencv.org/modules/ml/doc/normal_bayes_classifier.htmlRandom forests classifier
http://docs.opencv.org/modules/ml/doc/random_trees.htmlKNN classifier
http://docs.opencv.org/modules/ml/doc/k_nearest_neighbors.htmlShark Random forests classifier
http://image.diku.dk/shark/doxygen_pages/html/classshark_1_1_r_f_trainer.html.
It is noteworthy that training is parallel.Shark kmeans classifier
http://image.diku.dk/shark/sphinx_pages/build/html/rest_sources/tutorials/algorithms/kmeans.html
LibSVM classifier options¶
SVM Kernel Type -classifier.libsvm.k [linear|rbf|poly|sigmoid] Default value: linear
SVM Kernel Type.
Linear
Linear Kernel, no mapping is done, this is the fastest option.Gaussian radial basis function
This kernel is a good choice in most of the case. It is an exponential function of the euclidean distance between the vectors.Polynomial
Polynomial Kernel, the mapping is a polynomial function.Sigmoid
The kernel is a hyperbolic tangente function of the vectors.
SVM Model Type -classifier.libsvm.m [csvc|nusvc|oneclass] Default value: csvc
Type of SVM formulation.
C support vector classification
This formulation allows imperfect separation of classes. The penalty is set through the cost parameter C.Nu support vector classification
This formulation allows imperfect separation of classes. The penalty is set through the cost parameter Nu. As compared to C, Nu is harder to optimize, and may not be as fast.Distribution estimation (One Class SVM)
All the training data are from the same class, SVM builds a boundary that separates the class from the rest of the feature space.
Cost parameter C -classifier.libsvm.c float Default value: 1
SVM models have a cost parameter C (1 by default) to control the trade-off between training errors and forcing rigid margins.
Gamma parameter -classifier.libsvm.gamma float Default value: 1
Set gamma parameter in poly/rbf/sigmoid kernel function
Coefficient parameter -classifier.libsvm.coef0 float Default value: 0
Set coef0 parameter in poly/sigmoid kernel function
Degree parameter -classifier.libsvm.degree int Default value: 3
Set polynomial degree in poly kernel function
Cost parameter Nu -classifier.libsvm.nu float Default value: 0.5
Cost parameter Nu, in the range 0..1, the larger the value, the smoother the decision.
Parameters optimization -classifier.libsvm.opt bool Default value: false
SVM parameters optimization flag.
Probability estimation -classifier.libsvm.prob bool Default value: false
Probability estimation flag.
Boost classifier options¶
Boost Type -classifier.boost.t [discrete|real|logit|gentle] Default value: real
Type of Boosting algorithm.
Discrete AdaBoost
This procedure trains the classifiers on weighted versions of the training sample, giving higher weight to cases that are currently misclassified. This is done for a sequence of weighter samples, and then the final classifier is defined as a linear combination of the classifier from each stage.Real AdaBoost (technique using confidence-rated predictions and working well with categorical data)
Adaptation of the Discrete Adaboost algorithm with Real valueLogitBoost (technique producing good regression fits)
This procedure is an adaptive Newton algorithm for fitting an additive logistic regression model. Beware it can produce numeric instability.Gentle AdaBoost (technique setting less weight on outlier data points and, for that reason, being often good with regression data)
A modified version of the Real Adaboost algorithm, using Newton stepping rather than exact optimization at each step.
Weak count -classifier.boost.w int Default value: 100
The number of weak classifiers.
Weight Trim Rate -classifier.boost.r float Default value: 0.95
A threshold between 0 and 1 used to save computational time. Samples with summary weight <= (1 - weight_trim_rate) do not participate in the next iteration of training. Set this parameter to 0 to turn off this functionality.
Maximum depth of the tree -classifier.boost.m int Default value: 1
Maximum depth of the tree.
Decision Tree classifier options¶
Maximum depth of the tree -classifier.dt.max int Default value: 10
The training algorithm attempts to split each node while its depth is smaller than the maximum possible depth of the tree. The actual depth may be smaller if the other termination criteria are met, and/or if the tree is pruned.
Minimum number of samples in each node -classifier.dt.min int Default value: 10
If the number of samples in a node is smaller than this parameter, then this node will not be split.
Termination criteria for regression tree -classifier.dt.ra float Default value: 0.01
If all absolute differences between an estimated value in a node and the values of the train samples in this node are smaller than this regression accuracy parameter, then the node will not be split further.
Cluster possible values of a categorical variable into K <= cat clusters to find a suboptimal split -classifier.dt.cat int Default value: 10
Cluster possible values of a categorical variable into K <= cat clusters to find a suboptimal split.
Set Use1seRule flag to false -classifier.dt.r bool Default value: false
If true, then a pruning will be harsher. This will make a tree more compact and more resistant to the training data noise but a bit less accurate.
Set TruncatePrunedTree flag to false -classifier.dt.t bool Default value: false
If true, then pruned branches are physically removed from the tree.
Artificial Neural Network classifier options¶
Train Method Type -classifier.ann.t [back|reg] Default value: reg
Type of training method for the multilayer perceptron (MLP) neural network.
Back-propagation algorithm
Method to compute the gradient of the loss function and adjust weights in the network to optimize the result.Resilient Back-propagation algorithm
Almost the same as the Back-prop algorithm except that it does not take into account the magnitude of the partial derivative (coordinate of the gradient) but only its sign.
Number of neurons in each intermediate layer -classifier.ann.sizes string1 string2... Mandatory
The number of neurons in each intermediate layer (excluding input and output layers).
Neuron activation function type -classifier.ann.f [ident|sig|gau] Default value: sig
This function determine whether the output of the node is positive or not depending on the output of the transfer function.
Identity function
Symmetrical Sigmoid function
Gaussian function (Not completely supported)
Alpha parameter of the activation function -classifier.ann.a float Default value: 1
Alpha parameter of the activation function (used only with sigmoid and gaussian functions).
Beta parameter of the activation function -classifier.ann.b float Default value: 1
Beta parameter of the activation function (used only with sigmoid and gaussian functions).
Strength of the weight gradient term in the BACKPROP method -classifier.ann.bpdw float Default value: 0.1
Strength of the weight gradient term in the BACKPROP method. The recommended value is about 0.1.
Strength of the momentum term (the difference between weights on the 2 previous iterations) -classifier.ann.bpms float Default value: 0.1
Strength of the momentum term (the difference between weights on the 2 previous iterations). This parameter provides some inertia to smooth the random fluctuations of the weights. It can vary from 0 (the feature is disabled) to 1 and beyond. The value 0.1 or so is good enough.
Initial value Delta_0 of update-values Delta_{ij} in RPROP method -classifier.ann.rdw float Default value: 0.1
Initial value Delta_0 of update-values Delta_{ij} in RPROP method (default = 0.1).
Update-values lower limit Delta_{min} in RPROP method -classifier.ann.rdwm float Default value: 1e-07
Update-values lower limit Delta_{min} in RPROP method. It must be positive (default = 1e-7).
Termination criteria -classifier.ann.term [iter|eps|all] Default value: all
Termination criteria.
Maximum number of iterations
Set the number of iterations allowed to the network for its training. Training will stop regardless of the result when this number is reachedEpsilon
Training will focus on result and will stop once the precision isat most epsilonMax. iterations + Epsilon
Both termination criteria are used. Training stop at the first reached
Epsilon value used in the Termination criteria -classifier.ann.eps float Default value: 0.01
Epsilon value used in the Termination criteria.
Maximum number of iterations used in the Termination criteria -classifier.ann.iter int Default value: 1000
Maximum number of iterations used in the Termination criteria.
Random forests classifier options¶
Maximum depth of the tree -classifier.rf.max int Default value: 5
The depth of the tree. A low value will likely underfit and conversely a high value will likely overfit. The optimal value can be obtained using cross validation or other suitable methods.
Minimum number of samples in each node -classifier.rf.min int Default value: 10
If the number of samples in a node is smaller than this parameter, then the node will not be split. A reasonable value is a small percentage of the total data e.g. 1 percent.
Termination Criteria for regression tree -classifier.rf.ra float Default value: 0
If all absolute differences between an estimated value in a node and the values of the train samples in this node are smaller than this regression accuracy parameter, then the node will not be split.
Cluster possible values of a categorical variable into K <= cat clusters to find a suboptimal split -classifier.rf.cat int Default value: 10
Cluster possible values of a categorical variable into K <= cat clusters to find a suboptimal split.
Size of the randomly selected subset of features at each tree node -classifier.rf.var int Default value: 0
The size of the subset of features, randomly selected at each tree node, that are used to find the best split(s). If you set it to 0, then the size will be set to the square root of the total number of features.
Maximum number of trees in the forest -classifier.rf.nbtrees int Default value: 100
The maximum number of trees in the forest. Typically, the more trees you have, the better the accuracy. However, the improvement in accuracy generally diminishes and reaches an asymptote for a certain number of trees. Also to keep in mind, increasing the number of trees increases the prediction time linearly.
Sufficient accuracy (OOB error) -classifier.rf.acc float Default value: 0.01
Sufficient accuracy (OOB error).
KNN classifier options¶
Number of Neighbors -classifier.knn.k int Default value: 32
The number of neighbors to use.
Examples¶
From the command-line:
otbcli_TrainVectorClassifier -io.vd vectorData.shp -io.stats meanVar.xml -io.out svmModel.svm -feat perimeter area width -cfield predicted
From Python:
import otbApplication
app = otbApplication.Registry.CreateApplication("TrainVectorClassifier")
app.SetParameterStringList("io.vd", ['vectorData.shp'])
app.SetParameterString("io.stats", "meanVar.xml")
app.SetParameterString("io.out", "svmModel.svm")
app.SetParameterStringList("feat", ['perimeter', 'area', 'width'])
app.SetParameterString("cfield", "predicted")
app.ExecuteAndWriteOutput()