Logo

Support Vector Machine (SVM)

Used for solving both Classification (via [SVC]) and Regression (via [SVR]) problems.

Support Vector Machine (SVM)

Pasted image 20240904193641.png

"Margins" or Support Vectors are line parallel to Linear Model Line
The farther the distance [d] here, better the SVM

When there's multiple Linear Model Lines that can be made, SVM allows choosing the line with best fit

SVM Kernels

Take example below, SVM kernels acts on closely looped data points, where dataset is transformed into higher dimension to create separation

SVM Kernels runs transformation on data for better accuracy (better to use with GridSearchCV)

SVC on 2D -> 3D

Pasted image 20240905064954.png

SVC on 1D -> 2D

Pasted image 20240905065159.png

SVM Kernels Techniques

  • Polynomial

  • RBF

  • Sigmoid

Example: Using [Polynomial] Transformation

Before Transformation: Pasted image 20240905073931.png

After Transformation: Pasted image 20240905074035.png

Then normally, SVC Model training can be run for better accuracy

© 2025 All rights reservedBuilt with DataHub Cloud

Built with LogoDataHub Cloud