Analiza velikih mreža zasnovana na strojnom učenju

master's thesis

• Title (english): Machine Learning Based Large-Scale Network Analysis
• Author: Nikola Bukovac
• Mentor: Igor Čavrak (mentor)
• Committee member: Igor Čavrak (predsjednik povjerenstva)
• Committee member: Danko Basch (član povjerenstva)
• Committee member: Ivana Bosnić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2018-07-19, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Reprezentiranjem čvorova moguće je raditi analizu strojnim učenjem nad velikim
mrežama s obzirom na to da je svaki čvor prikazan kao vektor realnih vrijednosti određene veličine koja je mnogo manja od broja čvorova u mreži. Analiza je napravljena
korištenjem šest algoritama reprezentativnog učenja čvorova s različitim metodama
učenja kako bi se ustanovilo koje su metode najbolje. Izbor najboljeg algoritma za
reprezentiranje čvorova je ovisan i o metodi analize koju želimo napraviti nad skupom podataka, ali i o parametrima algoritma koji su odabrani pa je provedena i analiza
utjecaja parametara na utjecaj algoritama. Provedene su analize klasifikacijom čvorova, predviđanjem veza, grupiranjem čvorova i vizualizacije čvorova koje mogu imati
različita značenja ovisno o domeni primjene.

Abstract (english)

Using node representation learning to create node representations it is possible to
analyze large scale networks with machine learning because every node is representated as a vector of real numbers which size is a lot smaller than the number of nodes in
the network. Analysis was made using six representation learning algorithms with different learning methods to determine which methods were the best. The choice of the
best representation learning algorithm is dependent on the analysis method we want
to do on the dataset, but also on the chosen algorithm parameters and because of that
a parameter analysis is done. Analyzes were performed by classifying nodes, predicting links, grouping nodes, and visualizing nodes that could have different meanings
depending on the application domain.

• Keywords: velike mreže; strojno učenje; reprezentativno učenje čvorova; klasifikacija; vizualizacija; analiza
• Keywords (english): large scale networks; machine learning; node representation learning; classification; visualization; analysis
• Language: croatian
• URN:NBN: urn:nbn:hr:168:165477
• Study programme: Title: Computing; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2019-03-13 19:26:14