Fixed Points in SelfSimilar Analysis of Time Series.pdfVIP

8 9 9 Fixed Points in Self–Similar Analysis of Time 1 r Series a M 1 2 S. Gluzman and V.I. Yukalov 5 ] 1International Centre of Condensed Matter Physics h University of Brasilia, Brasilia, DF 70919–970, Brazil c e and m 2 Centre for Interdisciplinary Studies in Chemical Physics t- University of Western Ontario, London, Ontario N6A 3K7, Canada a t s . t a m - d n o Abstract c [ Two possible definitions of fixed points in the self–similar analysis 1 of time series are considered. One definition is based on the minimal– v 9 difference condition and another, on a simple averaging. From studying 5 stock market time series, one may conclude that these two definitions are 0 practically equivalent. A forecast is made for the stock market indices for 3 0 the end of March 1998. 8 9 / t a m - d n Time series analysis and forecasting have a long history and abundant litera- o ture, to mention just a few Refs. [1-3]. When analysing time series, one usually c : aims at constructing a particular model that could represent the available histor- v i ical data and, after such a model is constructed, one could use it for predicting X future. This kind of approach has been found rather reasonable for describing r a sufficiently smooth evolution, but it