Network Data Flow Clustering based on Unsupervised Learning
Not available until 9999-99-99
View/ Open
Use this link to cite
http://hdl.handle.net/2183/35398Collections
Metadata
Show full item recordTitle
Network Data Flow Clustering based on Unsupervised LearningAuthor(s)
Date
2019Citation
M. López-Vizcaíno, C. Dafonte, F. J. Novoa, D. Garabato, M. A. Álvarez and D. Fernández, "Network Data Flow Clustering based on Unsupervised Learning," 2019 IEEE 18th International Symposium on Network Computing and Applications (NCA), Cambridge, MA, USA, 2019, pp. 1-5, doi: 10.1109/NCA.2019.8935041.
Abstract
[Abstract]: Network communication data analysis is crucial in order to provide an adequate security level in computer infrastructures. As the volume of data and the number of features rise, the difficulties associated with its study also increase. A possible approach to this problem based on Artificial Intelligence techniques is to perform clustering, creating groups of similar characteristics. Self-Organized Maps are an artificial neural networks which perform dimensionality reduction as well as clustering over numerical data. In this paper, we propose a mixed numerical-categorical version of Self-Organized Maps algorithm applied to network communication data. Moreover a study of its enhancement with the inclusion of time dependent features is performed. The technique is tested with a well known dataset for Intrusion Detection Systems and it provides a technique to create different clusters or groups of different types of traffic by means of network information.
Keywords
Communication networks
Intrusion detection
Feature extraction
Computer security
Neural networks
Clustering algorithms
Self-Organizing Maps
IDS
Network Security
Categorical SOM
Visualization
Unsupervised Clustering
Intrusion detection
Feature extraction
Computer security
Neural networks
Clustering algorithms
Self-Organizing Maps
IDS
Network Security
Categorical SOM
Visualization
Unsupervised Clustering
Editor version
Rights
© 2019, IEEE