Spectral Analysis of Signals : The Missing Data Case

Spectral estimation is important in many fields including astronomy, meteorology, seismology, communications, economics, speech analysis, medical imaging, radar, and underwater acoustics. Most existing spectral estimation algorithms are devised for uniformly sampled complete-data sequences. However, the spectral estimation for data sequences with missing samples is also important in a wide range of applications.This lecture considers the spectral estimation problem in the case where some of the data samples are missing. The challenge is how to extend the existing spectral estimation techniques to deal with these missing-data samples. Recently, nonparametric adaptive filtering based techniques have been developed successfully for various missing-data spectral estimation problems. Collectively, these algorithms provide a comprehensive toolset for the missing-data problem based exclusively on the nonparametric adaptive filter-bank approaches. They provide the main topic of this lecture..The authors present the recently developed nonparametric adaptive filtering based algorithms for the missing-data case, namely gapped-data APES (GAPES) and the more general missing-data APES (MAPES).