International Association of Computer Science & Information Technology (IACSIT)
Hurst parameter estimation accuracy is very important in DDoS attack detection model based on traffic self-similarity changes. If the estimation accuracy is poor, the undetected and false detection should happen. This paper proposed the fractional fourier transform method to estimate hurst parameter. Combined with the idea of regression diagnosis and change point analysis, an effective, adaptive estimator of hurst parameter based on weighted least squares was presented. The authors used FGN data set to verify their idea.