Please use this identifier to cite or link to this item:
|Type:||Artigo de periódico|
|Title:||Robust Sequential Projection Algorithm: Teletraffic Applications|
Do Amaral W.C.
|Abstract:||An n-step ahead Robust Sequential Projection Algorithm - Robust SPA -, used for traffic forecast as well as for long term network planning is proposed. The algorithm uses the State Space Model with the state transition matrix being estimated by a relaxation technique. This lightly increases its complexity, as compared to a similar algorithm proposed in the literature, but renders the Robust SPA a better performance. The algorithm is also compared with the Expectation-Maximization technique and, in general, the Robust SPA yields a slightly poorer performance as far as the mean square error between projected and observed data is concerned. On the other hand, it is more robust and requires far less computational effort. Moreover, due to its simplicity, an automated version of the algorithm can be easily implemented for applications in which a great deal of data is to be forecasted.|
|Appears in Collections:||Unicamp - Artigos e Outros Documentos|
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.