What is the mathematical explanation behind a larger RBF kernel parameter resulting in a more linear decision boundary in SVM?

16 Views Asked by At

I know that as the kernel parameter for the RBF kernel increases, the Gaussian function becomes less peaked and broader. The reach of the points become larger meaning that farther datapoints have more weight. Intuitively, it makes sense that a larger reach means a smoother decision boundary. However, since I don't have an extensive background in mathematics, I find it difficult to come up with a more mathematical explanation. I was wondering whether anyone could help me with that?