ABSTRACT
Singular Spectrum Analysis (SSA) has widely and successfully been used in a number of different areas including biomedical signal processing, economics and finance, image processing, earth science and hydrology (for exam-
Signals
ple, see [1]–[16] and references therein). This method is particularly useful for analysing data with complex seasonal patterns and non-stationary trends and particularly in the cases where a single channel measurement is available. The SSA technique is a non-parametric method, and one of its advantages is that it can be used without the need for any assumptions, such as stationarity and normality of the data [1]. Initially, this method decomposes the time series into three components of trend, harmonics and noise. It then reconstructs the series, using the estimated trend and harmonic components and computes the forecasts based on the reconstructed series.
