A Granular Computing Approach for Multi-Labelled Sequences Classification in IEEE 802.11 Networks

Granato, Giuseppe; Martino, Alessio; Rizzi, Antonello

doi:10.1109/ijcnn55064.2022.9892473

Cited by 1 publication

(4 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Table 5 shows the results on the non-exclusive supervised problem, where both solutions achieve a Jaccard similarity score of ∼70%, with the best ones achieved for buffer length B ≥ 20. A similar phenomenon has been observed in [3], where a larger buffer length yields a larger amount of normal traffic, which constitutes the majority of instances in the dataset. The compressed training set cardinality plays a key role also in this case, with the GRALG classification system that is able to achieve interesting results with fewer examples, conversely to the GNN which requires a larger set of training graphs to learn from.…”

Section: Resultssupporting

confidence: 78%

“…Table 4 shows that, despite the difficulty to deal with such a complex problem, results are comparable with those obtained in [3]. This is particularly true if the granulated training set is big enough to bring a fair amount of information from the entire training graphs.…”

Section: Resultsmentioning

confidence: 51%

“…As GRALG is concerned, instead, the alphabet cardinality marks the dimensionality of the embedding space. If compared with our previous analysis (i.e., performed by buffering a sequence of packets with no transformation towards the graph domain [3]), we can see that the resulting alphabet is larger, yielding embedding space(s) with more than 120 features (i.e., information granules), after the feature selection step. This represents a physiological increase in terms of model complexity due to the deeper topological information brought by graphs.…”

Section: Resultsmentioning

confidence: 81%

“…To overcome this limitation, in the present work, we consider sequences of packets captured from the network traffic. In addition, however, rather than working on sequences of packets (as already explored in [3]), we derive graphs from ordered sequences of packets being transmitted over the network and use elements of those graphs to build an alphabet of symbols to encode the information carried in the traffic stream useful to classify attacks.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Graph-Based Multi-Label Classification for WiFi Network Traffic Analysis

et al. 2022

Self Cite

View full text Add to dashboard Cite

Network traffic analysis, and specifically anomaly and attack detection, call for sophisticated tools relying on a large number of features. Mathematical modeling is extremely difficult, given the ample variety of traffic patterns and the subtle and varied ways that malicious activity can be carried out in a network. We address this problem by exploiting data-driven modeling and computational intelligence techniques. Sequences of packets captured on the communication medium are considered, along with multi-label metadata. Graph-based modeling of the data are introduced, thus resorting to the powerful GRALG approach based on feature information granulation, identification of a representative alphabet, embedding and genetic optimization. The obtained classifier is evaluated both under accuracy and complexity for two different supervised problems and compared with state-of-the-art algorithms. We show that the proposed preprocessing strategy is able to describe higher level relations between data instances in the input domain, thus allowing the algorithms to suitably reconstruct the structure of the input domain itself. Furthermore, the considered Granular Computing approach is able to extract knowledge on multiple semantic levels, thus effectively describing anomalies as subgraphs-based symbols of the whole network graph, in a specific time interval. Interesting performances can thus be achieved in identifying network traffic patterns, in spite of the complexity of the considered traffic classes.

show abstract

Section: Resultssupporting

confidence: 78%

Section: Resultsmentioning

confidence: 51%

Section: Resultsmentioning

confidence: 81%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations