Paper Title
Classifying Diversified Attacks In Ids Using Data Mining Techniques

Abstract
Abstract—Latest information security techniques like user authentication, antivirus, firewalls, data encryption etc fail to prevent intrusion in any computer network. This may be attributed to the vulnerability in computer system or computer network. There is a need to use some sophisticated security tools like Intrusion Detection System (IDS) in order to protect our system/network. However these tools suffer from challenges posed by multi dimensionality in data. Data mining techniques like feature extraction algorithms based on mutual information are found to be effective to negotiate these technical challenges. The proposed approach measures mutual information through Information gain and then extracts features from bench marked data set KDDCUP99. Further the LSSVM classifier will detect the type of attack with better accuracy.