Generalised Singular Spectrum (GSS)

Updated 12 July 2026

Generalised Singular Spectrum (GSS) is a framework that extends classical SSA by embedding structured data into specialized singular-spectrum objects.
GSS adapts the standard embedding–decomposition–grouping–reconstruction process to multidimensional, functional, and non-Hermitian data for enhanced analysis.
In non-Hermitian applications, GSS analyzes complex band structures and edge modes via singular values and phase information, aiding signal extraction and filtering.

Searching arXiv for the cited papers on Generalised Singular Spectrum and related SSA extensions. Generalised Singular Spectrum (GSS) is a context-dependent term centered on singular-value-based structure extraction. In the SSA lineage, it is best understood as a broad generalization of classical Singular Spectrum Analysis in which the standard embedding–decomposition–grouping–reconstruction scheme is extended from univariate Hankel trajectory matrices to richer structured objects, including multivariate, multidimensional, shaped, functional, interval-valued, projected, and non-Hermitian constructions (Golyandina, 2019, Golyandina et al., 2013, Haghbin et al., 2019, Carvalho et al., 2020, Nicolsky et al., 2020). In a distinct 2025 usage, the term denotes the complex quantity

$\varepsilon_j(k)=\sigma_j(k)e^{-i\Arg \langle u_j(k),v_j(k)\rangle},$

derived from the singular value decomposition of a non-Hermitian dynamical matrix and interpreted as a band structure for driven-dissipative systems (Wanjura et al., 23 Sep 2025). The term is therefore not fully standardized, but in both lineages it preserves a common core: structured data are embedded into a singular-spectrum object, decomposed into elementary singular components, and then interpreted through reconstruction, filtering, or band topology.

1. Conceptual scope

Within the SSA family, the most general abstract formulation is a four-step scheme: embed the input object by an operator $T$ into a structured trajectory matrix from a set $H$ , decompose that trajectory matrix into rank-one elementary matrices, group them, and project grouped matrices back to $H$ before applying $T^{-1}$ (Golyandina, 2019). In this sense, “generalised singular spectrum” denotes not a single algorithm but a class of methods obtained by changing the embedding geometry, the structured matrix class, the decomposition operator, or the reconstruction rule while retaining the singular-spectrum logic.

The multidimensional/image literature makes this generalization particularly explicit. There, classical 1D SSA is extended in at least five senses: generalized embedding from 1D lag vectors to multidimensional moving-window patches, multidimensional lag-covariance PCA, adaptive filter-bank interpretation of eigenspaces, exact reconstruction through completeness of eigenvectors, and low-rank/noise separation by grouping or truncating components (Kume et al., 2015). This broader framing treats GSS less as a fixed named formalism than as a family resemblance across SSA extensions.

2. Classical SSA substrate

Classical SSA starts from a univariate series $X=(x_1,\dots,x_N)$ , chooses a window length $L$ , and forms the Hankel trajectory matrix $\mathbf X=\mathcal T(X)$ with $K=N-L+1$ columns. Its singular value decomposition is

$\mathbf X=\sum_{m=1}^d \sqrt{\lambda_m}\,U_mV_m^T,$

where $T$ 0 are the eigentriples, and grouped reconstruction is obtained by summing selected rank-one terms and projecting back to the Hankel class (Golyandina, 2019). In signal-extraction form, this can be written compactly as

$T$ 1

A complementary interpretation, foundational for many GSS variants, views SSA as a parallel FIR filter bank. The $T$ 2-th eigenvector defines an analysis filter

$T$ 3

and an overview filter

$T$ 4

with branch transfer function

$T$ 5

Because the branch impulse response is symmetric, each branch is zero-phase, and the full bank satisfies

$T$ 6

which is an exact-reconstruction partition of unity (Tomé et al., 2018). In the 1D filtering interpretation emphasized in the multidimensional SSA literature, each mode contributes the zero-phase transfer $T$ 7, with completeness

$T$ 8

so the singular spectrum is simultaneously an eigenspectrum, a filter bank, and a reconstruction system (Kume et al., 2015).

3. Generalizations within the SSA family

A first axis of generalization changes the embedding geometry. MSSA replaces one Hankel trajectory matrix by a stacked Hankel construction

$T$ 9

2D-SSA replaces it by a Hankel-block-Hankel trajectory matrix built from $H$ 0 moving windows, and Shaped 2D-SSA generalizes both the data support and the window support to arbitrary shapes $H$ 1 and $H$ 2, producing quasi-Hankel matrices and serving as a unifying implementation basis for MSSA and related extensions (Golyandina et al., 2013). In this framework, generalized singular-spectrum methods differ chiefly by the structure of $H$ 3.

A second axis changes the observation space itself. Functional SSA lifts the observations to the Hilbert space $H$ 4, embeds them into lagged vectors in $H$ 5, and replaces matrix SVD by the operator SVD

$H$ 6

of a compact trajectory operator $H$ 7 (Haghbin et al., 2019). Interval-Valued SSA instead embeds interval series into Hankel matrices of ordered pairs, uses a symbolic covariance matrix with entries

$H$ 8

reconstructs by interval diagonal averaging, and interprets grouped components as interval trendlines, cycles, or noise (Carvalho et al., 2020).

A third axis changes the decomposition operator or introduces prior structure. SSA with projection inserts row and column projectors $H$ 9 and $H$ 0 before residual SVD and is therefore semi-nonparametric; it is especially effective for polynomial trend extraction, particularly linear trends (Golyandina et al., 2015). Non-Hermitian SSA replaces the Hermitian covariance eigenproblem by the generalized pencil

$H$ 1

and produces transformed coordinates

$H$ 2

so that each retained coordinate is approximately a single exponential rather than an orthogonal variance-maximizing mixture (Nicolsky et al., 2020). A forecasting-oriented extension, called General SSA, leaves the decomposition stage unchanged but replaces the fixed linear recurrent formula by the state-dependent recurrence

$H$ 3

to address structural breaks (Rahmani et al., 2016).

4. Multidimensional filtering and image-domain GSS

The image-domain literature gives one of the clearest operational realizations of generalized singular spectrum. For an image

$H$ 4

an $H$ 5 moving window with $H$ 6 generates a trajectory matrix by sliding the window over the lattice. The eigenvectors of the lag-covariance matrix

$H$ 7

are arranged into adaptive 2D filters $H$ 8, and the decomposition is implemented as a two-step filtering system: forward filtering $H$ 9, reverse filtering by the point-symmetric filter $T^{-1}$ 0, and exact synthesis

$T^{-1}$ 1

In the Fourier domain, each component contributes a nonnegative zero-phase weighting, and completeness becomes

$T^{-1}$ 2

at every frequency (Kume et al., 2015).

For periodic images and rectangular windows, the lag-covariance matrix is bisymmetric: $T^{-1}$ 3 Hence it commutes with the exchange matrix $T^{-1}$ 4, and nondegenerate eigenvectors satisfy

$T^{-1}$ 5

For square windows this yields centrosymmetric or skew-centrosymmetric filters. In a $T^{-1}$ 6 Taylor expansion, symmetric filters cancel odd-order terms and behave like even-order differential operators, while antisymmetric filters cancel even-order terms and behave like odd-order differential operators. The dominant filter is therefore a smoother, intermediate filters act as directional edge enhancers or second-derivative detail operators, and small-eigenvalue filters behave as high-pass or noise filters. In the noisy $T^{-1}$ 7 Lenna example with an $T^{-1}$ 8 window, the RMS error

$T^{-1}$ 9

decreases until $X=(x_1,\dots,x_N)$ 0 and then increases, indicating that components $X=(x_1,\dots,x_N)$ 1 mainly reintroduce noise. This image-domain interpretation is one of the strongest arguments for viewing generalized SSA as an adaptive filter-bank theory rather than only as matrix factorization (Kume et al., 2015).

5. GSS as a non-Hermitian band structure

A 2025 work introduces the generalised singular spectrum as a specific complex band variable for driven-dissipative non-Hermitian systems. Starting from the Bloch dynamical matrix

$X=(x_1,\dots,x_N)$ 2

it defines

$X=(x_1,\dots,x_N)$ 3

The motivation is physical rather than purely algebraic: in such systems the scattering response

$X=(x_1,\dots,x_N)$ 4

is governed more directly by singular values and singular vectors than by the ordinary complex eigenvalue spectrum, especially in the presence of non-normality and the non-Hermitian skin effect. In the normal-matrix limit, the GSS reduces to the familiar band structure (Wanjura et al., 23 Sep 2025).

Within that framework, the GSS supports both point-gap and line-gap topology. The band-resolved point-gap invariant is

$X=(x_1,\dots,x_N)$ 5

which is also the winding of the $X=(x_1,\dots,x_N)$ 6-th GSS band around the origin. Under additional particle-hole symmetry of the doubled Hermitian matrix, one can define the $X=(x_1,\dots,x_N)$ 7-quantized line-gap invariant

$X=(x_1,\dots,x_N)$ 8

and the gap invariant

$X=(x_1,\dots,x_N)$ 9

The associated bulk-boundary correspondences distinguish two types of boundary modes: nonzero $L$ 0 yields singular zero modes with left and right singular vectors localized at opposite boundaries and thus directional amplification, whereas nonzero $L$ 1 yields finite-singular-value Hermitian-like edge modes with co-localized singular vectors and boundary-local response. The paper illustrates these two regimes with a 1D non-Hermitian SSH model and with a 2D non-Hermitian BBH-type model showing corner-to-corner amplification (Wanjura et al., 23 Sep 2025).

In the SSA literature, “Generalised Singular Spectrum” is not yet a settled proper noun. One survey of SSA extensions explicitly does not provide evidence that “GSS” is an established standard name synonymous with any one modification, and the image-decomposition work that is often read as strongly GSS-like does not formulate itself under that label (Golyandina, 2019, Kume et al., 2015). The term therefore functions more reliably as an umbrella description than as a universally fixed method name.

A further complication is acronym overload. In long-context language modeling, GSS denotes Gated State Spaces rather than Generalised Singular Spectrum (Torlak et al., 15 Jun 2026). In numerical linear algebra, GSS denotes generalized shift-splitting in saddle-point preconditioning (Salkuyeh et al., 2016). This suggests that any technical use of “GSS” requires an explicit local definition. Within the SSA family itself, the practical bottleneck is often no longer decomposition but grouping: automated identification methods based on low-frequency concentration and on angle regularity of singular-vector pairs have been proposed for SSA and extended to MSSA and 2D-SSA (Golyandina et al., 2023). A plausible implication is that, as singular-spectrum methods generalize across data types and operator geometries, automated grouping becomes part of the effective definition of a usable GSS pipeline.