In Chap. 2, the use of SSA for analyzing one-dimensional data is thoroughly examined. In this chapter, the use of models is minimal so that the main techniques can be considered as non-parametric and descriptory. Relations with algorithms of space rotation and many other methods aiming at achieving better separability of signal from noise are outlined. The common problems of smoothing, filtration, and splitting of a time series into identifiable components such as trend, seasonality, and noise are thoroughly discussed and illustrated on case studies with real data. An important issue of automatization of the SSA methods is also considered in detail.
CITATION STYLE
Golyandina, N., Korobeynikov, A., & Zhigljavsky, A. (2018). SSA Analysis of One-Dimensional Time Series (pp. 31–120). https://doi.org/10.1007/978-3-662-57380-8_2
Mendeley helps you to discover research relevant for your work.