Support Vector Machines are a new type of universal learning machines
that implement SRM inductive inference. These machines can construct
approximations of multi-dimensional functions in various sets of
functions: splines, Fourier expansions, radial basis function
expansion etc.  SV machines effectively control the trade off between
the accuracy of approximation and the complexity of the approximating
function.
 
In this talk we will discuss practical implementation issues of SV
machines for large data bases, consider examples of function
approximation in different sets of functions and demonstrate the
effectiveness of controlling complexity versus accuracy of an
approximation.

We will also discuss some examples of solving real-life problems using
SV machines.