ABSTRACT
In this chapter a range of useful techniques, other than those used in the earlier chapters, are discussed. These range from basic statistical approaches to artificial neural networks (ANNs) which have been applied with great effect to some monitoring problems. The discussion highlights strengths and weaknesses, and then suggests possible future aims. The text then covers a range of other techniques, some of which have been extensively applied, such as cepstrum analysis and empirical mode decomposition, whilst others such as cyclostationary analysis are relatively new. The range discussed also covers the Hilbert transform (HT), time–frequency and Wigner–Ville, wavelet analysis, higher-order spectra, and kernel density estimation. The relevance of all these techniques is outlined.
