Adaptive Real-time Anomaly-based Intrusion Detection using Data Mining and Machine Learning Techniques

Die zunehmende Vernetzung der Informations- und Kommunikationssysteme führt zu einer weiteren Erhöhung der Komplexität und damit auch zu einer weiteren Zunahme von Sicherheitslücken. Klassische Schutzmechanismen wie Firewall-Systeme und Anti-Malware-Lösungen bieten schon lange keinen Schutz mehr vor Eindringversuchen in IT-Infrastrukturen. Als ein sehr wirkungsvolles Instrument zum Schutz gegenüber Cyber-Attacken haben sich hierbei die Intrusion Detection Systeme (IDS) etabliert. Solche Systeme sammeln und analysieren Informationen von Netzwerkkomponenten und Rechnern, um ungewöhnliches Verhalten und Sicherheitsverletzungen automatisiert festzustellen. Während signatur-basierte Ansätze nur bereits bekannte Angriffsmuster detektieren können, sind anomalie-basierte IDS auch in der Lage, neue bisher unbekannte Angriffe (Zero-Day-Attacks) frühzeitig zu erkennen. Das Kernproblem von Intrusion Detection Systeme besteht jedoch in der optimalen Verarbeitung der gewaltigen Netzdaten und der Entwicklung eines in Echtzeit arbeitenden adaptiven Erkennungsmodells. Um diese Herausforderungen lösen zu können, stellt diese Dissertation ein Framework bereit, das aus zwei Hauptteilen besteht. Der erste Teil, OptiFilter genannt, verwendet ein dynamisches "Queuing Concept", um die zahlreich anfallenden Netzdaten weiter zu verarbeiten, baut fortlaufend Netzverbindungen auf, und exportiert strukturierte Input-Daten für das IDS. Den zweiten Teil stellt ein adaptiver Klassifikator dar, der ein Klassifikator-Modell basierend auf "Enhanced Growing Hierarchical Self Organizing Map" (EGHSOM), ein Modell für Netzwerk Normalzustand (NNB) und ein "Update Model" umfasst. In dem OptiFilter werden Tcpdump und SNMP traps benutzt, um die Netzwerkpakete und Hostereignisse fortlaufend zu aggregieren. Diese aggregierten Netzwerkpackete und Hostereignisse werden weiter analysiert und in Verbindungsvektoren umgewandelt. Zur Verbesserung der Erkennungsrate des adaptiven Klassifikators wird das künstliche neuronale Netz GHSOM intensiv untersucht und wesentlich weiterentwickelt. In dieser Dissertation werden unterschiedliche Ansätze vorgeschlagen und diskutiert. So wird eine classification-confidence margin threshold definiert, um die unbekannten bösartigen Verbindungen aufzudecken, die Stabilität der Wachstumstopologie durch neuartige Ansätze für die Initialisierung der Gewichtvektoren und durch die Stärkung der Winner Neuronen erhöht, und ein selbst-adaptives Verfahren eingeführt, um das Modell ständig aktualisieren zu können. Darüber hinaus besteht die Hauptaufgabe des NNB-Modells in der weiteren Untersuchung der erkannten unbekannten Verbindungen von der EGHSOM und der Überprüfung, ob sie normal sind. Jedoch, ändern sich die Netzverkehrsdaten wegen des Concept drif Phänomens ständig, was in Echtzeit zur Erzeugung nicht stationärer Netzdaten führt. Dieses Phänomen wird von dem Update-Modell besser kontrolliert. Das EGHSOM-Modell kann die neuen Anomalien effektiv erkennen und das NNB-Model passt die Änderungen in Netzdaten optimal an. Bei den experimentellen Untersuchungen hat das Framework erfolgversprechende Ergebnisse gezeigt. Im ersten Experiment wurde das Framework in Offline-Betriebsmodus evaluiert. Der OptiFilter wurde mit offline-, synthetischen- und realistischen Daten ausgewertet. Der adaptive Klassifikator wurde mit dem 10-Fold Cross Validation Verfahren evaluiert, um dessen Genauigkeit abzuschätzen. Im zweiten Experiment wurde das Framework auf einer 1 bis 10 GB Netzwerkstrecke installiert und im Online-Betriebsmodus in Echtzeit ausgewertet. Der OptiFilter hat erfolgreich die gewaltige Menge von Netzdaten in die strukturierten Verbindungsvektoren umgewandelt und der adaptive Klassifikator hat sie präzise klassifiziert. Die Vergleichsstudie zwischen dem entwickelten Framework und anderen bekannten IDS-Ansätzen zeigt, dass der vorgeschlagene IDSFramework alle anderen Ansätze übertrifft. Dies lässt sich auf folgende Kernpunkte zurückführen: Bearbeitung der gesammelten Netzdaten, Erreichung der besten Performanz (wie die Gesamtgenauigkeit), Detektieren unbekannter Verbindungen und Entwicklung des in Echtzeit arbeitenden Erkennungsmodells von Eindringversuchen.

Conceptual knowledge discovery and data analysis

In this paper, we discuss Conceptual Knowledge Discovery in Databases (CKDD) in its connection with Data Analysis. Our approach is based on Formal Concept Analysis, a mathematical theory which has been developed and proven useful during the last 20 years. Formal Concept Analysis has led to a theory of conceptual information systems which has been applied by using the management system TOSCANA in a wide range of domains. In this paper, we use such an application in database marketing to demonstrate how methods and procedures of CKDD can be applied in Data Analysis. In particular, we show the interplay and integration of data mining and data analysis techniques based on Formal Concept Analysis. The main concern of this paper is to explain how the transition from data to knowledge can be supported by a TOSCANA system. To clarify the transition steps we discuss their correspondence to the five levels of knowledge representation established by R. Brachman and to the steps of empirically grounded theory building proposed by A. Strauss and J. Corbin.

Dual retrieval in conceptual information systems

Conceptual Information Systems provide a multi-dimensional conceptually structured view on data stored in relational databases. On restricting the expressiveness of the retrieval language, they allow the visualization of sets of realted queries in conceptual hierarchies, hence supporting the search of something one does not have a precise description, but only a vague idea of. Information Retrieval is considered as the process of finding specific objects (documents etc.) out of a large set of objects which fit to some description. In some data analysis and knowledge discovery applications, the dual task is of interest: The analyst needs to determine, for a subset of objects, a description for this subset. In this paper we discuss how Conceptual Information Systems can be extended to support also the second task.

Efficient data mining based on formal concept analysis

Formal Concept Analysis is an unsupervised learning technique for conceptual clustering. We introduce the notion of iceberg concept lattices and show their use in Knowledge Discovery in Databases (KDD). Iceberg lattices are designed for analyzing very large databases. In particular they serve as a condensed representation of frequent patterns as known from association rule mining. In order to show the interplay between Formal Concept Analysis and association rule mining, we discuss the algorithm TITANIC. We show that iceberg concept lattices are a starting point for computing condensed sets of association rules without loss of information, and are a visualization method for the resulting rules.

Conceptual knowledge discovery with frequent concept lattices

Knowledge discovery support environments include beside classical data analysis tools also data mining tools. For supporting both kinds of tools, a unified knowledge representation is needed. We show that concept lattices which are used as knowledge representation in Conceptual Information Systems can also be used for structuring the results of mining association rules. Vice versa, we use ideas of association rules for reducing the complexity of the visualization of Conceptual Information Systems.

Evolving Classifiers - Evolutionary Algorithms in Data Mining

Data mining means to summarize information from large amounts of raw data. It is one of the key technologies in many areas of economy, science, administration and the internet. In this report we introduce an approach for utilizing evolutionary algorithms to breed fuzzy classifier systems. This approach was exercised as part of a structured procedure by the students Achler, Göb and Voigtmann as contribution to the 2006 Data-Mining-Cup contest, yielding encouragingly positive results.

Fast computation of concept lattices using data mining techniques

We present a new algorithm called TITANIC for computing concept lattices. It is based on data mining techniques for computing frequent itemsets. The algorithm is experimentally evaluated and compared with B. Ganter's Next-Closure algorithm.

Semantic web mining and the representation, analysis, and evolution of web space

Semantic Web Mining aims at combining the two fast-developing research areas Semantic Web and Web Mining. This survey analyzes the convergence of trends from both areas: Growing numbers of researchers work on improving the results of Web Mining by exploiting semantic structures in the Web, and they use Web Mining techniques for building the Semantic Web. Last but not least, these techniques can be used for mining the Semantic Web itself. The second aim of this paper is to use these concepts to circumscribe what Web space is, what it represents and how it can be represented and analyzed. This is used to sketch the role that Semantic Web Mining and the software agents and human agents involved in it can play in the evolution of Web space.

Hierarchies of conceptual scales

Formal Concept Analysis allows to derive conceptual hierarchies from data tables. Formal Concept Analysis is applied in various domains, e.g., data analysis, information retrieval, and knowledge discovery in databases. In order to deal with increasing sizes of the data tables (and to allow more complex data structures than just binary attributes), conceputal scales habe been developed. They are considered as metadata which structure the data conceptually. But in large applications, the number of conceptual scales increases as well. Techniques are needed which support the navigation of the user also on this meta-level of conceptual scales. In this paper, we attack this problem by extending the set of scales by hierarchically ordered higher level scales and by introducing a visualization technique called nested scaling. We extend the two-level architecture of Formal Concept Analysis (the data table plus one level of conceptual scales) to many-level architecture with a cascading system of conceptual scales. The approach also allows to use representation techniques of Formal Concept Analysis for the visualization of thesauri and ontologies.

Network properties of folksonomies

Social resource sharing systems like YouTube and del.icio.us have acquired a large number of users within the last few years. They provide rich resources for data analysis, information retrieval, and knowledge discovery applications. A first step towards this end is to gain better insights into content and structure of these systems. In this paper, we will analyse the main network characteristics of two of the systems. We consider their underlying data structures – socalled folksonomies – as tri-partite hypergraphs, and adapt classical network measures like characteristic path length and clustering coefficient to them. Subsequently, we introduce a network of tag co-occurrence and investigate some of its statistical properties, focusing on correlations in node connectivity and pointing out features that reflect emergent semantics within the folksonomy. We show that simple statistical indicators unambiguously spot non-social behavior such as spam.

Network properties of folksonomies

Social resource sharing systems like YouTube and del.icio.us have acquired a large number of users within the last few years. They provide rich resources for data analysis, information retrieval, and knowledge discovery applications. A first step towards this end is to gain better insights into content and structure of these systems. In this paper, we will analyse the main network characteristics of two of these systems. We consider their underlying data structures – so-called folksonomies – as tri-partite hypergraphs, and adapt classical network measures like characteristic path length and clustering coefficient to them. Subsequently, we introduce a network of tag cooccurrence and investigate some of its statistical properties, focusing on correlations in node connectivity and pointing out features that reflect emergent semantics within the folksonomy. We show that simple statistical indicators unambiguously spot non-social behavior such as spam.