The Use of Artificial Intelligence for the Intrusion Detection System in Computer Networks

摘要

We discuss the application of Artificial Intelligence for the design of intrusion detection systems (IDS) applied on computer networks. For this purpose, we use J48 rand Clonal-G [5] immune artificial system Algorithms, in WEKA software, with the purpose to classify and predict intrusions in KDD-Cup 1999 and Kyoto 2006 databases. We obtain for the KDD-Cup 1999 database 92.69% for ClonalG and 99.91% of precision for J48 respectively. For the Kyoto University 2006 database, we obtain 95.2% for ClonalG and 99.25% of precision for J48. Finally, based on these results we propose a model to detect intrusions using AI techniques. The main contribution of the paper is the adaptability of the CLONAL-G Algorithm and the reduction of database attributes by using Genetic Search.