Presentation 2007-03-06
Suppression of boundary effect and introduction of scale correlation for Wavelet based traffic prediction
Naoya MATSUSUE, Hiroshi HASEGAWA, Ken-ichi SATO,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In this paper, we propose a method of traffic prediction based on undecimated Haar Wavelet transform and vector autoregressive model. Firstly we show that so called Boundary condition, that means some future inputs are required to derive current Wavelet coefficients, can be resolved by combining undecimated formulation and Haar Wavelet. Unfortunately, localization in frequency domain is poor for the Haar Wavelet and it results in strong correlation between Wavelet coefficients at different frequencies is strongly correlated. Next, therefore, we propose a prediction method that employs vector autoregressive model so as to introduce the correlation. Numerical experiment shows that the proposed method has smaller prediction error than a conventional method [Luan '05] based on A Trous transform, one of undecimated transforms, and Recursive Least-Squares algorithm to each frequency coefficients.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Long range dependence / Wavelet transform / Scale correlation / Vector autoregressive model / Traffic prediction
Paper # PN2006-81
Date of Issue

Conference Information
Committee PN
Conference Date 2007/2/27(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Photonic Network (PN)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Suppression of boundary effect and introduction of scale correlation for Wavelet based traffic prediction
Sub Title (in English)
Keyword(1) Long range dependence
Keyword(2) Wavelet transform
Keyword(3) Scale correlation
Keyword(4) Vector autoregressive model
Keyword(5) Traffic prediction
1st Author's Name Naoya MATSUSUE
1st Author's Affiliation Graduart School of Electrical Engineering and Computer Science, Nagoya University()
2nd Author's Name Hiroshi HASEGAWA
2nd Author's Affiliation Graduart School of Electrical Engineering and Computer Science, Nagoya University
3rd Author's Name Ken-ichi SATO
3rd Author's Affiliation Graduart School of Electrical Engineering and Computer Science, Nagoya University
Date 2007-03-06
Paper # PN2006-81
Volume (vol) vol.106
Number (no) 546
Page pp.pp.-
#Pages 4
Date of Issue