[Truncated] We began by using wavelets as a tool to model the dynamics of nonlinear dynamical systems. Due to the nature of dynamical systems data the standard wavelet transforms were not effective, and so a method was developed which fitted dictionaries of wavelet functions to the data, which incorporated an optimization step to ensure goodness of fit. Following this we investigated modelling the distribution of data points in order to estimate the average mutual information of a system. Again, standard procedures proved unsuitable for our purposes, and a new method was developed from tomographic principles.
|Qualification||Doctor of Philosophy|
|Publication status||Unpublished - 2002|