A0614
Title: A novel environmental system-focused empirical mode decomposition analysis: Application to Minas passage
Authors: Ian Kenny - The Open University (United Kingdom) [presenting]
Dhouha Kbaier - The Open University (United Kingdom)
Abstract: The application of Empirical Mode Decomposition (EMD) to four components of the Minas Passage data is considered. Specifically, we present a finding which places the number of Intrinsic Mode Functions (IMFs) as a function of the number of inputs, rather than the number of observations $N$. The previous empirical finding that the number of IMF required was not more than $\log2N$ meant that the number of functions required grew with the number of observations. In addition, being able to relate the number of functions required to the number of inputs has the effect of enabling the identification of the systemic relationships within the signal, this is achieved through the judicious application of Time-Dependent Intrinsic Correlation (TDIC) to the identified IMFs. The discovery of an input-related empirical measure for the number of IMFs also addresses the issue of mode mixing, by viewing the number of modes in relation to the system rather than the number of observations. Therefore, the smaller signal components are treated as if they are artefacts of the interactions between the signals. By applying this method, we have been able to show that EMD can be used as a starting point for building system-focused adaptive models
