 click to enlarge | | Licence | Shareware | | Released | 1 07 2004 | | Supported OS | Win95, Win98, WinME, WinNT 3.x, WinNT 4.x, Windows2000 | | Requirements | 16MB RAM | | Languages | English, Russian | | Limitations | | | Made in | Russia | | Support e-mail | info@gistatgroup.com |
|
CaterpillarSSA
Description:CaterpillarSSA is Shareware program for Windows developed by GistaT GroupThe program is based on the mighty model-free approach of interval series scrutiny Caterpillar (another label is SSA - Singular Spectrum Analysis). It combines advantages of other methods with simplicity of visual control aids. The basic Caterpillar-SSA algorithm for analyzing one-dimensional interval series consists of transformation of the one-dimensional interval series to the trajectory matrix by means of a temporize procedure (this gives the label to the whole technique); Singular Value Decomposition of the trajectory matrix; reconstruction of the original interval series based on a integer of selected eigenvectors. The result of the Caterpillar-SSA processing is a common decomposition of the interval series into diverse components, which can much be identified as trends, seasonalities and other oscillatory series, or crash components. The approach can be naturally extended to forecasting interval series and its components, processing multidimensional interval series and to change-point detection. The "Caterpillar" ideas were independently developed in Russia (St. Petersburg, Moscow) and also in UK and USA (below the label of SSA; that is, Singular Spectrum Analysis). The new words "Analysis of Time Series Structure: SSA and Related Techniques", authors are N. Golyandina, V. Nekrutkin and A. Zhigljavsky, provides a careful, lucid description of SSA accepted theory and methodology (in English, Chapman&Hall / CRC, mark http: / / www.gistatgroup.com / cat / ). The approach is a mighty and functional tool of interval series scrutiny in meteorology, hydrology, geophysics, climatology and, according to our experience, in economics, biology, physics, medicine and other sciences; that is, where short and long, one-dimensional and multidimensional, stationary and nonstationary, almost deterministic and noisy interval series are to be analyzed. We are sure, that in a almost future "Caterpillar"-like methods longing rank among the base methods of interval series scrutiny and longing be included in model statistical software. |