Adaptive filters
Noisy signal processing methods

Determining the Wiener filter for a noisy stationary signal, as described in the previous paragraph, requires knowledge of the second-order moments of the "signal, noise" pair. In reality, this information is often unavailable: an adaptive method can be envisaged, based, for example, on iterative optimization such as stochastic gradient or recursive least squares. Note that such adaptive methods are frequently used in non-stationary signal processing. Here again, we restrict ourselves to the case of real, discrete, monovariable, scalar quantities, bearing in mind that these structures generalize to multivariable, vector cases.