Input data preprocessing for bots detection using the dendritic cells algorithm

摘要

The dendritic cells algorithm (DCA) is an immune-inspired intelligent algorithm and is based on an abstract model of the behavior of dendritic cells (DCs) in immunology. Meanwhile, a number of bots based on diverse protocols, including P2P, IRC, DNS and HTTP, have become an increasing threat to network security. So many researchers use the DCA to detect a single bot in computer system. But the DCA's input is the time series data consisting of antigens and signals. The complex raw data must be preprocessed to form the input stream of antigens and signals. The data preprocessing of the DCA includes raw data collection phase, selection, extraction and mapping phase of antigens and signals. This paper proposed the several methods of input data preprocessing for bots detection using the DCA. These methods are suit for the different bots detection, such as IRC and P2P bots detection.