Support Vector (SVM)

A Support Vector Machine is a supervised algorithm that can classify cases by finding a

separator.
1. Mapping the data to a high-dimensional feature space
2. Finding a separator

How do we transfer data in such a way that a separator could be drawn as a hyper plane? 
How can we find the best or optimized hyper plane separator after transformation?

Data transformation

We usually choose different functions in turn and compare the results.

Using SVM to find the hyper plane

One reasonable choice as the best hyper plane is

the one that represents 
the largest separation or margin between the two
classes. 
So the goal is to choose a hyper plane with as big a
margin as possible.