I have a function that is continuous and differentiable over $\mathbb{R}$ and its support is the whole real line.
I want to approximate it through a linear combination of Gaussian functions. I know that approximation through RBF is a common thing in Neural Networks, but I'm not familiar myself with this field.
I would like to know if there is any tool in R that I can use to generate an approximation, being able to retrieve each of the centers, deviances and weights used. It would be also important to be able indicate in which norm do we try to minimize the approximation error.
Best regards and thanks in advance
I was learning RBF 3 within 3 weeks ago . I had many problem in first step. But I find an example ,and do it for myself many time ,then I understand how to use Rbf's .Now I think ,this way may help you .IF not I do apologize !$$ $$