N.N. Kyklin, T.S. Lazebnyk, K.A. Sevostyanova

Annotations languages:

This article examines the alternative approach, used for analysis and prediction markets, which is a singular spectral analysis of the SSA (Singular Spectrum Analysis). The idea behind this approach is based on dynamic modification of the method of principal components. Using the method of singular value decomposition of time series can be distinguished from alternating series of separate components such as trend, periodic components and random noise. Carried out the decomposition of the time series into components using singular value decomposition of the matrix scan. Depend not only on the components of time series, but the class received time series.