Table Of Contents
Train SVM Classifier (G Dataflow)
Sets the classifier session in to use the SVM classifier engine and configures the SVM parameters to use.
classifier session in
Reference to the classifier session on which the node operates.
kernel options
Cluster of parameters to configure the kernel.
kernel
Kernel that the classifier uses to determine if a sample is a texture or contains a defect.
The SVM classifier is a linear classifier. Use a nonlinear kernel to transform samples with nonlinear feature information to a dimension where the feature information is linearly separable.
| Name | Description |
|---|---|
| linear | Applies a linear kernel to the sample. Use this kernel If the number of features per sample is high. |
| polynomial | Applies a polynomial kernel to the sample. |
| Gaussian | Applies a Gaussian kernel to the sample. |
| RBF | (Default) Applies a radial basis function (RBF) kernel to the sample. |
degree
Degree of the polynomial kernel.
The polynomial kernel becomes a linear kernel if you specify a degree of 1.
gamma
Gamma value for the polynomial and RBF kernels.
A high value requires more support vectors to classify the sample. Use a high value for samples with regularly distributed feature information, and a low value for samples with irregularly distributed feature information.
coefficient
Coefficient of the polynomial kernel.
sigma
Sigma value for the Gaussian kernel.
A higher value produces a smoother Gaussian function and fewer support vectors. Use a high value for samples with regularly distributed feature information, and a low value for samples with irregularly distributed feature information.
Default: 1
cache(MB)
Cache size, in megabytes, for kernel operations.
model options
Cluster of parameter to configures the SVM model.
SVM model
SVM model that trains the classifier.
| Name | Value | Description |
|---|---|---|
| C-SVC | 0 | The C-SVC model allows the SVM algorithm to clearly separate samples that are separated by a very narrow margin. If the SVM algorithm cannot define a clear margin, it uses the cost parameter to allow some training errors and produce a soft margin. If the cost value is too high it prohibits training errors, producing a narrow margin and rigid classification. |
| nu-SVC | 1 | In the nu-SVC model, the nu parameter controls training errors and the number of support vectors. The nu value specifies both the maximum ratio of training errors and the minimum number of support vectors relative to the number of samples. Nu must be greater than 0 and cannot exceed 1. A higher nu value increases tolerance for variation in the texture, but may also increase tolerance for texture defects. If nu is too high, training produces too many training errors to be useful. |
| one class | 2 | (Default) In the one-class model, the SVM algorithm considers the spatial distribution information for each sample to determine whether the sample belongs to the known class. |
tolerance
Maximum gradient of the quadratic function used to compute support vectors.
Default: 0.001
max. iteration
Maximum number of iterations.
nu
Value of nu.
Values range from 0.1 to 1. The default value is 0.1 A higher nu value increases tolerance for variation in the texture, but may also increase tolerance for defects. If the texture classifier does not perform as expected because the trained texture samples do not represent every possible variation of the texture, try increasing the value of nu.
shrinking
Classifier that uses shrinking heuristics to attempt to reduce the number of variables involved in the classification computation.
Shrinking may reduce processing time if the number of iterations is large.
cost
Penalty for training errors.
If the cost value is too high it prohibits training errors, producing a narrow margin and rigid classification. Decrease the cost value to allow more training errors and produce a softer margin between classes.
weighted cost
Cost weight for each label.
label
Label associated with the weight coefficient.
weight
Weighted coefficient for label.
error in
Error conditions that occur before this node runs.
The node responds to this input according to standard error behavior.
Standard Error Behavior
Many nodes provide an error in input and an error out output so that the node can respond to and communicate errors that occur while code is running. The value of error in specifies whether an error occurred before the node runs. Most nodes respond to values of error in in a standard, predictable way.
Default: No error
classifier session out
Reference to the classifier session the node creates.
training results
Array of statistical information for each class in the classifier.
class
Class into which the classifier session categorizes the input sample.
standard deviation
Standard deviation from the mean of all samples in class.
number of samples
Number of samples in class.
number of SVs
Number of support vectors in class.
class distance table
Table giving the mean distance from each class to each other class.
error out
Error information.
The node produces this output according to standard error behavior.
Standard Error Behavior
Many nodes provide an error in input and an error out output so that the node can respond to and communicate errors that occur while code is running. The value of error in specifies whether an error occurred before the node runs. Most nodes respond to values of error in in a standard, predictable way.
Where This Node Can Run:
Desktop OS: Windows
FPGA: Not supported
Web Server: Not supported in VIs that run in a web application