ResearchTrend.AI
  • Papers
  • Communities
  • Events
  • Blog
  • Pricing
Papers
Communities
Social Events
Terms and Conditions
Pricing
Parameter LabParameter LabTwitterGitHubLinkedInBlueskyYoutube

© 2025 ResearchTrend.AI, All rights reserved.

  1. Home
  2. Papers
  3. 2404.16583
29
0

Fast Machine-Precision Spectral Likelihoods for Stationary Time Series

25 April 2024
Christopher J. Geoga
ArXivPDFHTML
Abstract

We provide in this work an algorithm for approximating a very broad class of symmetric Toeplitz matrices to machine precision in O(nlog⁡n)\mathcal{O}(n \log n)O(nlogn) time with applications to fitting time series models. In particular, for a symmetric Toeplitz matrix Σ\mathbf{\Sigma}Σ with values Σj,k=h∣j−k∣=∫−1/21/2e2πi∣j−k∣ωS(ω)dω\mathbf{\Sigma}_{j,k} = h_{|j-k|} = \int_{-1/2}^{1/2} e^{2 \pi i |j-k| \omega} S(\omega) \mathrm{d} \omegaΣj,k​=h∣j−k∣​=∫−1/21/2​e2πi∣j−k∣ωS(ω)dω where S(ω)S(\omega)S(ω) is piecewise smooth, we give an approximation FΣFH≈D+UVH\mathbf{\mathcal{F}} \mathbf{\Sigma} \mathbf{\mathcal{F}}^H \approx \mathbf{D} + \mathbf{U} \mathbf{V}^HFΣFH≈D+UVH, where F\mathbf{\mathcal{F}}F is the DFT matrix, D\mathbf{D}D is diagonal, and the matrices U\mathbf{U}U and V\mathbf{V}V are in Cn×r\mathbb{C}^{n \times r}Cn×r with r≪nr \ll nr≪n. Studying these matrices in the context of time series, we offer a theoretical explanation of this structure and connect it to existing spectral-domain approximation frameworks. We then give a complete discussion of the numerical method for assembling the approximation and demonstrate its efficiency for improving Whittle-type likelihood approximations, including dramatic examples where a correction of rank r=2r = 2r=2 to the standard Whittle approximation increases the accuracy from 333 to 141414 digits for a matrix Σ∈R105×105\mathbf{\Sigma} \in \mathbb{R}^{10^5 \times 10^5}Σ∈R105×105. The method and analysis of this work applies well beyond time series analysis, providing an algorithm for extremely accurate direct solves with a wide variety of symmetric Toeplitz matrices. The analysis employed here largely depends on asymptotic expansions of oscillatory integrals, and also provides a new perspective on when existing spectral-domain approximation methods for Gaussian log-likelihoods can be particularly problematic.

View on arXiv
Comments on this paper