Core idea:
Use a feature function to transform original data from to some other
Basic example of this is to transform a Linear Classifier through the origin to one not through the origin.
Perceptron through origin transformation
We can even turn un-separable datasets into separable ones!
XOR data set transformation example
Original XOR dataset (1D version) does not have a linear separator:
Using the transformation :
A linear separator in the space is non-linear in the original space.
- For example, can be a separator in the space, with the half-plane labeled as positive.
- The corresponding separator in the original space can then be found by considering what points are on the boundary of – obviously, the answer is and ;
Visualizing the the original separator:
-
This is a very useful and generalizable strategy – it serves as the basis for kernel methods.
-
One systematic strategy for constructing a new feature space is to use a polynomial basis.
-
Features can also be improved through Feature Engineering.