Support Vector Machine (SVM)

1. Overview

2. Description

2.1 Components

The line that separates these two classes of points. And at the same time, it has the maximum margin which means this distance so this line is drawn equidistant. And that's margin's So the sum of these two distances has to be maximized in order for this line to be the result of the SVM.

And these two points are actually called the support vectors. So basically they these two points are supporting this whole algorithm. So even if you get rid of all the rest of the points that thing will change the algorithm will be exactly the same.

So these other points don't contribute to the result of the algorithm only these two points are contributing and therefore they called the supporting vectors you can call them supporting points but in reality, they are vectors. And this is why because in a multi-dimensional space when you have more than just two variables you can have three-five 10 or 100 variables.

2.1.1 Margin

2.1.2 Maximum margin classifier

2.1.3 Soft Margin

2.2 Specialists

So imagine you're trying to teach a machine how to distinguish between apples and oranges how to classify a fruit into either an apple an orange.

A machine would try to learn from the apples that are very like apples so it would know what an apple is. And it also tried to learn from oranges so it would know what an orange actually is and that's how most of the machine learning algorithms work and then based on that it would be able to come up with some predictions and classifying four new data elements and variables that you would get it in the case of support vector machine.

Instead of looking at the most stocks standard apples and stocks and oranges what this support victualling

machines do is they actually look at the apples that are very much like an orange so here you can see an apple which is not your standard Apple is orange and color right. So it's very easy to infuse this apple of an orange and they would look at oranges which are not stock standard oranges which are more like apples than anything else so you can order the Lemon here. So those of us in the image just out of the oranges the SVM would pick the one that is that looks the most like an apple in this case. We have a green orange. It's not normal to have a green orange when you think of orange. And so what that is is. Those are the support vectors so the support vectors you can see that they're actually very close to the boundary so they're very close to the apple or the red one would be very close to the green ones and the orange or the green mark here would be very close to the red ones and therefore the support vector machine in that sense you can think of it is like a more extreme type of algorithm a very rebellious type of algorithm a very risky type of algorithm because it looks at a very extreme case which is very close to the boundary and it uses that to construct its analysis. And that in itself makes the support vector machine algorithms very special very different from most of the other machine learning algorithms. And that's why at times they perform much better than non-supported of vector machine algorithms.

3. Reference