Nonparametric modeling and spatiotemporal dynamical systems

M. Abel

Home/Archive/arXiv

arXiv · nlin/0202058

Nonparametric modeling and spatiotemporal dynamical systems

M. Abel

View original source ↗Read source document ↗Export metadata ↓

Abstract

In this article, it is described how to use statistical data analysis to obtain models directly from data. The focus is put on finding nonlinearities within a generalized additive model. These models are found by the means of backfitting algorithms or more general versions, like the alternating conditional expectation value method. The method is illustrated by numerically generated data. As an application the example of vortex ripple dynamics, a highly complex fluid-granular system is treated.

Explore related subjects

nlin.PS nlin.CD

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

M. Abel. 2002-08-14. Nonparametric modeling and spatiotemporal dynamical systems. https://arxiv.org/abs/nlin/0202058

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

Topological defect trains excited by a boundary discontinuity in the two-dimensional complex Ginzburg-Landau equation

With numerical simulations of the complex Ginzburg-Landau equation in a two-dimensional space, we find that a stationary but non-zero Dirichlet boundary condition with a discontinuity can drive a periodic train of evenly spaced topological defects that drift along and in proximity to the boundary.

nlin.PS↗

arXiv2026

Generation of discrete breathers from linear to nonlinear Anderson disorder in electrical lattices

We investigate linear and nonlinear localization in a disordered electrical lattice composed of coupled LC resonators. In the linear regime, the normal-mode spectrum and spatial voltage profiles are characterized theoretically and experimentally, revealing strongly localized Anderson modes induced by capacitive disorder. Good agreement is found between numerical predictions and experimental measurements of resonance frequencies, spatial profiles, inverse participation ratios, and localization lengths. Nonlinearity is introduced through varactor diodes, allowing the numerical continuation of a selected Anderson-localized mode into a family of discrete breathers. Floquet analysis indicates that these nonlinear localized solutions remain spectrally stable over the investigated frequency range, while their spatial localization evolves with increasing amplitude. Experimental measurements reveal nonlinear resonance bending, frequency shifts, and hysteresis, in qualitative agreement with the predicted nonlinear behavior. These results provide numerical evidence for the emergence of stable, disorder-pinned discrete breathers from Anderson-localized modes, supported by experimental observations of the associated nonlinear resonant response, and highlight the interplay between disorder and nonlinearity in electrical lattices.

nlin.PS↗

arXiv2026

Series approximation of discrete-soliton solutions of the DNLS equation

We study standing waves of the discrete nonlinear Schrödinger (DNLS) equation by means of perturbative series expansions in the inter-site coupling parameter $\varepsilon$. Power-series approximations are derived for three representative configurations: a single-site soliton and the two two-site ones, namely the in-phase and the out-of-phase configurations. Although higher-order truncations are increasingly accurate for small $\varepsilon$, the expansions possess a small radius of convergence and cease to be useful well before $\varepsilon={\cal O}(1)$. To extend their range of validity we apply the Euler convergence-acceleration method, constructing an optimally transformed series for each lattice site individually. The transformed solution is shown to have substantially better convergence radius and preserves the localization of the solution much better than the original series solution. We then insert the series profiles into the linearization matrix of the DNLS equation in order to obtain improved approximations of the stability eigenvalues, with particular attention to the unstable real eigenvalue of the in-phase configuration. Optimal truncation of the corresponding expansion yields a closed-form ${\cal O}(\varepsilon^{5/2})$ estimate, which significantly extends the known ${\cal O}(\varepsilon^{1/2})$ leading-order result to considerably larger couplings, while the semi-numerical use of the accelerated profiles reproduces the numerically computed eigenvalue to values of $\varepsilon$ of ${\cal O}(1)$. Throughout, both the solution profiles and their stability eigenvalues are benchmarked against the numerical solution obtained through a continuation procedure from the anti-continuum limit.

nlin.PS↗