The wavelet transform was introduced in the 19S0's and it was developed as an alternative in tin short time Fourier transform. The wavelets theory is very popular in signal processing and pattern recognition and its applications are still growing. This paper presents the wavelet transform in nonparametric regression. The use of wavelets, in statistical applications was pioneered by D. Donoho and I. Johnstone. Here we discuss their methodology - wavelet shrinkage. The wavelet transform is compared with another nonparametric regression method - splines.Transformacja falkowa została zaproponowana na początku lat osiemdziesiątych, jako alternatywa do transformacji Fouriera. Metoda ta szybko znalazła swoje zastosowanie w teorii sygnałów oraz w r...