Narrow Results

In order to solve the problem of detection efficiency and the detection speed in botnet detection, a novel botnet detection method is proposed based on hill-climbing algorithm and FARIMA. At first, the evaluation indexes are presented in this method, and botnet and infection hosts are quickly searched with hill-climbing algorithm. Then, FARIMA model is introduced to cut down the long-correlation of detection index. Finally, a simulation was conducted to research on the key factors with MATLAB. The result shows that, compared to other algorithms, it has good adaptability, and it could effectively search for infected hosts and botnets.

A method is developed for the automatic detection of the onset of scaling for long-range dependent (LRD) time series and other asymptotically scale-invariant processes. Based on wavelet techniques, it provides the lower cutoff scale for the regression that yields the scaling exponent. The method detects the onset of scaling through the dramatic improvement of a goodness-of-fit statistic taken as a function of this lower cutoff scale. It relies on qualitative features of the goodness-of-fit statistic and on features of the wavelet analysis. The method is easy to implement, appropriate for large data sets and highly robust. It is tested against 34 time series models and found to perform very well. Examples involving telecommunications data are presented.

The FARIMA models, which have long-range-dependence (LRD), are widely used in many areas. Through the derivation of a precise characterization of the spectrum and variance time function, we show that this family is very atypical among LRD processes, being extremely close to the fractional Gaussian noise in a precise sense which results in ultra-fast convergence to fGn under rescaling. Furthermore, we show that this closeness property is not robust to additive noise. We argue that the use of FARIMA, and more generally fractionally differenced time series, should be reassessed in some contexts, in particular when convergence rate under rescaling is important and noise is expected.