SMOOS — A program for the filtration of non-stationary statistical series

✍ Scribed by V.B. Zlokazov

• Publisher: Elsevier Science
• Year: 1984
• Tongue: English
• Weight: 43 KB
• Volume: 35
• Category: Article
• ISSN: 0010-4655
• DOI: 10.1016/s0010-4655(84)82826-x

✦ Synopsis

A method is described for constructing numerical high and low frequency filters for the filtration of the trajectories of strongly non-stationary stochastic processes (e.g. with a trend of the type of resonance functions). Measures of function oscillations and function variability are introduced, and by making use of them the problem of constructing the abovementioned filters is formulated in terms of the calculus of variations. A compact algorithm for the numerical implementation of the method is given and the corresponding standard FORTRAN-subroutines are described. A comparison with other Filters is carried out.