-Yash Dodwani

Introduction

Support Vector Machine is a supervised learning algorithm which identifies the best hyperplane to divide the dataset, at its core SVM tries to find the best 2-D line or Hyper-plane( in case of higher dimensions ) that separates the different classes of Data points. The best separator is the one which maximizes the margin (distance between the separator and the nearest data point from each class). The nearest data points are called support vectors ,hence the name . It can be used for both Classification and regression , but most popularly used for classification. It has high accuracy and computation speed ( depending on the size of the data ) of the model.

Some of the imp Terms

• Support Vectors : Points which are closest to the hyperplane
• Hyperplane : Subspace with the dimension one lower than it’s ambient space. It serves to divide the space into multiple sections.
• Margin: the distance between the hyperplane and support vectors from either points
• Kernel : A mathematical function used to transform input data into different forms

Types of SVM

Linear SVM

when the data is perfectly linearly separable only then we can use Linear SVM . Perfectly linear separable means hat the data points can be classified into two classes using single straight line .

Non-Linear SVM

When the data is not linearly separable then we can use Non-Linear SVM which means the data points can’t be divided into two classes using a single line ( if 2D ) , then we use some advanced techniques like kernel tricks to classify them

How Does SVM works

Suppose we have a data set consisting of two classes Blue and Green . Now we want to classify new data point as either blue or green .

To classify these points we can have different decision boundaries , so which one of these boundaries we will choose? (note: since our plane is 2D here , so our decision boundary will be a straight line , in case of higher dimensions it will be a hyperplane)

So the best hyperplane will be the one which has maximum distance from the support vector from each side, or one with largest margin, and this is the main aim of SVM.