ML model visualizer

SVM Classifier

Find the decision boundary that maximizes the margin between classes.

TypeClassificationModelMaximum MarginKernelLinearOutputClass Label
1.0
0.15

Decision Space

-6-4-20246-6-4-20246Feature 1Feature 2
Class AClass BDecision boundary
Decision boundaryNot trainedSupport vectors-Training accuracy-Training points48

How SVM Works

  1. 1
    Map training pointsPlot labeled observations in feature space.
  2. 2
    Test separating boundariesCompare candidate class boundaries.
  3. 3
    Maximize the marginChoose the widest reliable separation.
  4. 4
    Identify support vectorsKeep the points that define the margin.

SVM Classification

  1. Map labeled examples into the feature space
  2. Search for a boundary that separates the classes
  3. Maximize the margin around the decision boundary
  4. Identify the support vectors closest to the margin
  5. Use the boundary sign to classify new points

Model Metrics

Accuracy-Support Vectors-Margin Width-Margin Violations-C Parameter1.0KernelLinear

Support Vectors

#ClassCoordinatesScore
Advance to the support-vector step to inspect them.

Kernel Explorer

Linear kernels create a straight boundary or trend.