Singular Spectrum Analysis

The theory and algorithms of SSA and its extensions, which are used in BioSSA, can be found in the following papers supported by the NG13-083 grant of Dynasty Foundation:

Singular spectrum analysis as a method of time series analysis has well-elaborated theory and solves various problems: time series decomposition, trend extraction, periodicity detection and extraction, signal extraction, denoising, filtering, forecasting, missing data imputation, change point detection, spectral analysis among them. Since the method does not need a model given a priori, it is called nonparametric and is well suited for exploratory analysis of time series.

SSA was originally proposed for time series, but was extended to 2D digital images. 2D-SSA and related subspace based methods find applications in texture analysis, seismology, spatial gene expression data, medical imaging, etc., and are gaining increasing popularity.

The SSA and 2D-SSA methods deal with series and rectangular images and have subseries and rectangles as moving windows. This can limit applications of the methods; for example, the methods hardly process circular-shaped images, images with gaps and so on. Due to this reason, Shaped 2D-SSA can be proposed, which can deal with images of arbitrary shape and arbitrary window shapes.

General scheme of SSA-like methods is as follows:

Scheme of SSA-like methods