Measuring of reliability of network anomalies detection using Methods of discrete wavelet analysis

For the detection of the anomalies in the computer and telecommunication networks in “online” mode we propose the method, based on the discrete wavelet decomposition of the traffic and the algorithms of statistical processing, based on the F - test. We use two sliding windows with two threshold values that allow us to guarantee the high efficiency of the detection of the network anomalies´ bursts. We have shown that using of our developed algorithm by using discrete wavelet analysis gives the strong detection of anomalies simultaneously on several levels of the decomposition.