SVM(Basics)

SVM( support vector machine). You may be thinking it’s hard, but it’s easier than you think. Let’s go ahead! Before We moving ahead, We need to understand 5 basic concepts of SVM.

1. Hyperplanes.
2. Marginal Distance.
3. Support Vectors.
4. Linear Separable Line.
5. Non-Linear Separable Line.

HYPERPLANES:

In linear or logistic regression we used to separate the two features by using the linearly separable line. yes, that’s it, if you’re separating more than 2 features in your data that line is called hyperplanes. yes, you got it!

Hyperplane.

Marginal Distance:

It’s nothing but in linear and logistic regression we draw a hyperplane simply but here we are adding another function, name in sub-hyperplanes, confused! wait. Nothing but you need to draw two parallel lines near to the Hyperplanes with the condition of touch the first data on both sides. And Calculating both the distance is called Marginal Distance.

Hint: the higher the marginal distance the model will perform better.

The doted parallel lines are sub-hyperplanes and the distance between the hyperplane and sub-hyperplane is called Marginal Distance.

Support Vectors:

Support vectors are nothing, the sub-hyperplane was drawn by some condition( whenever find the first data from the hyperplanes, just draw the parallel plane). The data which touching the sub-hyperplanes that data are considered to be Support Vectors.

Linear Separable line:

• Nothing but if you’re separating the two features with the line(same as linear and logistic) that is called the linear separable line.
• But in many cases, problems are not that much easy to separate like this. For that, we use separately another function. which we will see later.

NON-linear Separable line:

• It’s is opposite to a linear separable line. we can’t separate easily we use a separate function for this, which we will talk about later.

non-linear line.

CONCLUSION:

• We have all covered the basics of SVM. I hope it gives some info about SVM.