You are on page 1of 4

Kernel Trick for Hyperplane

Why we need kernel hyper parameter?

Suppose we have points on x-axis. Now how we will differentiate circle from square?

In order to separate circles from squares we will introduce a y-axis and apply a
mathematical function.
Let y=x^2
How Kernel Works?
• Now to classify non-linear data, kernel introduces new dimension and make it
easy to separate things.
• Kernel apply mathematical operation to provided input and classify them using
new dimensions.
• The choice of mathematical operation is decided as per need.
Example

2-D
3-D
Types of Kernels in SVM

NAME FORMULA
LINEAR KERNEL
POLYNOMIAL KERNEL
HOMOGENIOUS KERNEL
GAUSSION KERNEL
SIGMOID KERNEL

You might also like