Method of Constructing a Nonlinear Approximating Scheme of a Complex Signal: Application Pattern Recognition

A method for identification of structures a complex signal and noise suppression based on nonlinear approximating schemes is proposed. When we do not know the probability distribution signal, problem identifying its can be solved by constructing adaptive in an orthonormal basis. The mapping constructed applying threshold functions, parameters which noisy data are estimated to minimize risk. In absence priori information about useful presence high level, use optimal ineffective. paper introduces threshold, assessed basis posterior Application natural has confirmed effectiveness.