Deconvolution problems in nonparametric statistics (Q951135)

This book gives an introduction to deconvolution problems in nonparametric statistics. It covers the following topics: Density estimation based on contaminated data, nonparametric regression with errors-in-variables and image and signal reconstruction. The author intends to provide a comprehensive overview on results derived during the last twenty years and to give a discussion on modern and recently solved problems. Estimation methods are introduced and the underlying ideas are explained, the asymptotic theory and adaptive choice of smoothing parameters are derived and the applicability of the considered methods is underlined by some real data examples. Basic knowledge in analysis, algebra, elementary probability theory and measure theory is necessary to understand the proofs of the theorems. Since Fourier analysis plays an important role in deconvolution, basic tools of this field are provided in a separate chapter. The main target group of readers are scientists and graduate students working in mathematical statistics. However, the book is also interesting for people working in econometrics, biometrics, and other fields of applied statistics.