Support Vector Machines
Large Margin Intuition
Using an SVM
if Gaussian Kernel, need to choose sigma^2
Not all similarity functions make valid kernels(need to satisfy technical condition called "Mercer's Theorem" to make sure SVM packages' optimizations run correctly, and do not diverge).
More esoteric: String kernel, chi-square kernel, histogram intersection kernel
K SVMs, one to distinguish one from the rest.