Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Exponential prediction models based on sequence operators

: Sydow, A.; Lin, Y.; Liu, S.F.; Denu, R.; Mennell, W.


Kybernetes 30 (2001), Nr.4, S.397-410
ISSN: 0368-492X
Fraunhofer FIRST ()

One of the difficulties experienced in applications of exponential prediction models of time series is resolved by introducing sequence operators. This approach is different from all known modelings of time series and regression analysis. Also, discontinuities (or outliers), appearing in a given time series, can be naturally absorbed by applying sequence operators. This end is important since, in terms of predictions, outliers or discontinuities in the data might signal major changes in the pattern of the time series data.