0% found this document useful (0 votes)
6 views3 pages

SVM 1 SummaryNotes

support vector machines
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
6 views3 pages

SVM 1 SummaryNotes

support vector machines
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

SVM-Introduction

- Support Vector Machine (SVM) is a powerful supervised machine learning


algorithm.
- It works by finding the optimal hyperplane that best separates data points into
different classes while maximizing the margin between the classes, making it
highly effective in high-dimensional spaces.

What is the key idea behind SVM?


𝑇
- The best hyperplane π: 𝑤 𝑥 + 𝑏 = 0 classifying between 2 classes is
the one that has maximum gap/margin (d) between itself and the closest
+ve and -ve datapoints.

Margin/gap
The hyperplane parallel to π, that touches the closest +ve point:
+ 𝑇
π : 𝑤 𝑥 + 𝑏 = 1;
The hyperplane parallel to π, that touches the closest -ve point:
− 𝑇
π :𝑤 𝑥 + 𝑏 = − 1
+ − 2
Margin is measured as the distance between them: 𝑑(π , π ) = ||𝑤||
;
- where w is the weight of the model.
The optimization problem in SVM
2
Optimization problem: 𝑚𝑎𝑥 ||𝑤||

The goal is to maximize generalization.

Support Vectors
These are the data points:-
- Are within the margin
- Or, are misclassified
+ −
- Or, which lies on the hyperplanes (π , π )
- ∝𝑖 = 0 for nonsupport vectors, whereas, ∝𝑖 > 0 for support vectors.

Different Types of SVM model


A. Hard Margin SVM:
- The simplest form of the SVM model.
- It assumes no data point can lie inside the Margin.
- Rarely works in real-life problems.

B. Soft Margin SVM:


- Introduces ζ as error for an incorrectly placed data point.
- ζ = 0 ; datapoint is placed such that the hyperplane
classifies the point correctly
- ζ > 1 ; datapoint is placed such that the hyperplane
classifies the point incorrectly
- ζ < 1 ; datapoint is placed such that the hyperplane still
manages to classify the point correctly, but lies inside the
margin.
- Linear soft margin SVM is similar to LogReg

You might also like