Κατεύθυνση / ειδίκευση Δικτύωση Υπολογιστών (ΔΙΚ)
Library of the School of Science

2013-08-09

2013

Βλάχος Χριστόφορος

Αθανασία Αλωνιστιώτη Αναπλ. Καθηγήτρια

A distributed data mining framework for knowledge extraction and dissemination in future internet environments

English

Κατανεμημένο πλαίσιο εξόρυξης δεδομένων για την εξαγωγή και διάχυση γνώσης μη επεξεργασμένων δικτυακών δεδομένων πάνω σε αρχιτεκτονικές μελλοντικών διαδικτύων

The last years, Network operators are struggling to perform efficient network
management due to the prominent increase of data that surpass their
infrastructures, the increased user demands and the complexity that the
different technologies introduce. Those issues require novel data mining
approaches able to process and exploit the voluminous data, enabling different
SON mechanisms to act upon the network infrastructures. To this end, we
introduce a Distributed Data Mining Framework (DDMF), which is exploiting
cutting-edge data mining tools such as spectral clustering and SVMs; based on a
novel NM approach called UMF, DDMF extracts knowledge from raw network and
service data and reveals the hidden context from massive datasets. To our
knowledge, this is the first work that introduces such sophisticated data
mining techniques for network management. The validation of DDMF is performed
in a WLAN environment for a Load identification scenario and the results prove
the efficiency of our work in comparison with a previous approach.

Network management, Data mining, Knowledge extraction, Future internet, Machine learning

Yes

8-10

Yes

21

37

File access is restricted.

https://pergamos.lib.uoa.gr/uoa/dl/object/1320811