Radial Basis Functions
- Instead of using a linear function, we try to learn a function of the form
where each is an instance from where the kernel function
is defined so that it decreases as the distance
increases.
- is a user-defined constant that specifies the number
of kernel functions to be included.
- Note that, even thought is a global approx to
, the contribution from each of the terms is
localized to a region near .
- It is common to choose to be a Gaussian centered around
- It has been shown that this can approximate
any function with arbitrarily small error, given enough
and given that each can be separately
specified.