The decomposition of a multivariate signal is an important tool for the analysis of measured or simulated data leading to possible detection of the relevant subspace or the sources of the signal. A new method - dynamical component analysis (DyCA) - is based on modeling the signal by a set of coupled ordinary differential equations. Its derivation and its features are presented in-depth. The corresponding algorithm is nearly as simple as principal component analysis (PCA). The results obtained by DyCA however yield a deeper insight into the underlying dynamics of the data. To illustrate the broad area of possible applications a set of examples of analyzing data by DyCA is presented - involving both measured EEG, motion and ECG data as well as data obtained from stochastic differential equations. Thereby our alternative tool for dimensionality reduction is compared to results obtained PCA and ICA and demonstrate the gain of this approach.
29 November 2018 in Frontiers in Applied Mathematics and Statistics
Dynamical Systems Based Modeling (DSBM) is a method to decompose a multivariate signal leading to both a dimensionality reduction and parameter estimation describing the dynamics of the signal. We present this method and its application to EEG data sets of Petit-Mal epilepsies considering Shilnikov chaos as the underlying dynamic interaction. We demonstrate the power of this method compared to conventional decomposition methods like PCA and ICA. Since the fitting quality showed a strong correlation to the ictal phases of the signal, we performed a cross validation on seizure detection with a resulting specifity of 84% and sensitivity of 75%. By applying DSBM in a moving window setup we investigated the comparability of the obtained dynamic models and tested the hypothesis of Shilnikov chaos in terms of linear stability analysis for each of the investigated windows. Thereby we could corroborate the Shilnikov hypothesis for approx. 50% of the relevant windows.
A signal decomposition is presented that disentangles the deterministic and stochastic components of a multivariate time series. The dynamical component analysis (DyCA) algorithm is based on the assumption that an unknown set of ordinary differential equations (ODEs) describes the dynamics of the deterministic part of the signal. The algorithm is thoroughly derived and accompanied by a link to the GitHub repository containing the algorithm. The method was applied to both simulated and real-world data sets and compared to the results of principal component analysis (PCA), independent component analysis (ICA), and dynamic mode decomposition (DMD). The results demonstrate that DyCA is capable of separating the deterministic and stochastic components of the signal. Furthermore, the algorithm is able to estimate the number of linear and non-linear differential equations and to extract the corresponding amplitudes. The results demonstrate that DyCA is an effective tool for signal decomposition and dimension reduction of multivariate time series. In this regard, DyCA outperforms PCA and ICA and is on par or slightly superior to the DMD algorithm in terms of performance.