A new Data mining Based Approach for Network

Intrusion Detection

Submitted by
Abstracts
• In Information Security, intrusion detection is the act of
detecting actions that attempt to compromise the
confidentiality, integrity or availability of a resource.
• Intrusion detection does not, in general, include
prevention of intrusions.
• In this paper, we are mostly focused on data mining
techniques that are being used for such purposes.
• We debate on the advantages and disadvantages of
these techniques.
• Finally we present a new idea on how data mining can
aid IDSs.
Existing System
• Distributed Intrusion Detection System

• Emarald

• The MINDS system

• ISOA-Internet Security Officer’s Assistance.

Drawbacks…
• Cannot detect Unknown Attacks

• False Positive

• False Negative

• Data Overload
Proposed System
• Anomaly Detection
• Misuse or Signature based Detection.
• Remove the normal activity from the attacks.
• Identifying the false alarm signal
• Find the anomalies attack
• Identify long, ongoing patterns.
Technologies used
• Data summarization
• Data Visualization
• Association of the rule discovery
• Classification
Algorithms
• Binary Classification.
• Machine Learning
• Feature selection
• Genetic Algorithms