Initialize SVM

ERDAS IMAGINE Help

HGD_Variant
16.5
HGD_Product
ERDAS IMAGINE
HGD_Portfolio_Suite
Producer

Category: Classification

Associated Namespace: IMAGINE

Default

Show All Ports

InitializeSVM_Default.PNG

InitializeSVM_AllPorts

Description

The operator defines and trains a Support Vector Machine (SVM) classifier which is used as an input for classifying data using the Classify Using Machine Intellect operator. All non-geometry attribute fields in the TrainingData feature data (except for ClassAttributeName) will be used for training. The Select Attributes or Remove Attributes operator can be used to tailor the training data's attribute schema before the training data is used in the Initialize SVM operator.

Support Vector Machine (SVM) is a supervised machine learning algorithm that performs classification by finding optimal hyper planes that separates the classes. The minimum distance between a hyper plane and a class is called a Margin. The optimal hyper plane is the one which has the maximum margin.

The training data points that lie near the dividing hyper plane, and if removed would change the position of the hyper plane, are called support vectors.

Limitations

None

Connections

Name

Objects Supported

Description

Shown by Default

Required

Default Behavior or Behavior if not Required

TrainingData

IMAGINE.Features

The data to be used for defining and training the classifier

checkmark_gray

checkmark_gray

ClassAttributeName

IMAGINE.String

The attribute field from TrainingData which defines the classes into which the resulting MachineIntellect may classify features.

checkmark_gray

checkmark_gray

RandomSeed

IMAGINE.Unsigned

Seed for the random number generator that is used during the training. Fixing the seed at a specific value allows repeated executions of the operator with the same input data to produce the exact same results.

Empty. The trained SVM classifier is random. That is, running this operator multiple times will generate (slightly) different results.

MachineIntellect

IMAGINE.MachineIntellect

The trained SVM classifier

checkmark_gray

Related Operators

Initialize CART, Initialize K-Nearest Neighbors, Initialize Naive Bayes, Initialize Random Forest, Classify Using Machine Intellect

Example Model

InitializeSVM_example