Portal de Periódicos da CAPES

Software expedites singular‐spectrum analysis of noisy time series

1995; Wiley; Volume: 76; Issue: 2 Linguagem: Inglês

10.1029/eo076i002p00012

2324-9250

Michael D. Dettinger, Michael Ghil, Christopher M. Strong, William Weibel, Pascal Yiou,

Statistical and numerical algorithms

A Singular‐Spectrum Analysis (SSA) Toolkit is now available for use on computers with X‐windows capabilities. The Toolkit provides free, compact, and easy access to SSA and several powerful spectral analysis techniques not readily available in common statistical packages. The current Toolkit, Version 1.1, provides tools for breaking down univariate, possibly nonstationary time series into trends, periodic oscillations, other statistically significant components, and noise; reconstructing the contributions of selected components to the time series; and estimating power spectra of the time series or any of its components by modern spectral methods.