Abstract:
|
Singular Spectrum Analysis (SSA) is a nonparametric tecnique for signal extraction in time series based on principal components. After the estimation of the principal components SSA identifies the frequencies associated to them extract the related signals. We propose a new variant of SSA, Circulant SSA (CSSA) tha, on the contrary, allows to automatically identify the principal component associated to any previously chosen frequency. We also prove the validity of CSSA for the nonstationary case. Through several sets of simulations, we show the good properties of our approach: it is reliable, fast, automatic and more accurate than other SSA versions. Finally, we apply Circulant SSA to the Industrial Production Index of six countries. The analysis of the cycles estimated are in accordance with the dated recessions from the OECD showing the reliability of the proposed procedure for business cycle analysis. Also, we use CSSA to deseasonalize and analyze the strong separability of the estimated components.
|