A New Model of Anomaly Network Flow Detection
Tu Xu · 2005
The main methods to detect anomaly network flow are through comparison between the system's actual ac- tion pattern and normal action pattern. There is a trade-off between real-time and veracity in anomaly network flow de- tection, this paper presents a new model -adjustable Chi-Square T~2 model, according to a set of trusty coefficient. Through adjusting these coefficients, a trusty measure can contribute more value to the total anomaly value with a high factor, this will observably improve detection rate.