- [Home]
- [Research achievement]
- [Research achievement detail]
Title | A Note on HTTP Traffic Analysis of the Time Series Model with a Time Varying Density Parameter (in Japanese) |
---|---|
Authors | Daiki Koizumi 、Toshiyasu Matsushima 、Shigeichi Hirasawa |
Released Year | 2005 |
Format | Conference |
Category | Others |
Jounal Name | Proceedings of the 28th Symposium on Information Theory and its Applications |
Jounal Page | pp.729-732 |
Published Year | 2005 |
Published Month | 11 |
Abstract (English) |
In Internet traffic analysis, there are some approaches assuming the probabilistic distribution on the request arrivals. One of the classical approaches has been touse the time series model with the stationary Poisson distribution. The stationary Poisson distribution, however, can be insufficient to express Internet (heavily burst) traffic. In this paper, we would deal with a time series model with the time varying (nonstationary) Poisson distribution for HTTP traffic analysis. In our model, the density parameterof the Poisson distribution is time varying. Furthermore, our model has some features in terms of theoretical guarantee of estimated traffic value as well as computational advantage. Finally, we would show some analysis results of the real HTTP traffic data under our model. |
Note (English) |
1 |
Manuscript | |
Presentation |