ABSTRACT
This conclusion presents some closing thoughts on the concepts covered in the preceding chapters of this book. The book demonstrates that wavelets or other multiscale analyses can be constructed on a wide variety of data. Spline methods and local polynomial smoothing approaches can be lifted higher to a multiscale version. The book discusses how to construct a wavelet analyis of irregular point sets from scratch. Possible domains of further multiscale research include graphical and network data, neural networks. In functional data, each observation is a different function, possibly showing discontinuities at different locations, for which adaptive non-equispaced multiscale analyses could be developed. Images are a classical, nice, and visual illustration in a text on wavelets.
