Analiza velikih skupova podataka primjenom algoritama dubinske analize

master's thesis

• Title (english): Big Data Analytics by Using Data Mining Algorithms
• Author: Nikola Medaković
• Mentor: Boris Vrdoljak (mentor)
• Committee member: Boris Vrdoljak (predsjednik povjerenstva)
• Committee member: Igor Mekterović (član povjerenstva)
• Committee member: Marko Banek (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2014-07-07, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Sve veća količina pohranjivanih podataka donosi sa sobom nove izazove kada se govori o njihovoj analizi zbog njihove rastuće količine, brzine generiranja i raznolikosti. Twitter je društvena mreža čiji su podaci o interakcijama korisnika primjer Velikog skupa podataka. Kroz njega je lako pratiti proces propagacije informacija kroz mrežu, što uvelike olakšava proučavanje društvenog utjecaja korisnika. Društveni utjecaj se u ovom kontekstu može definirati kao sposobnost korisnika da svojim djelovanjem utječe na druge korisnike. U ovom radu su se dohvatili podaci o korisnicima i njihovim objavama na temu grafičkih kartica, a iz njih su se generirale mreže sljedbenika, spominjanja i ponovnog objavljivanja. Te mreže su se predale algoritmima Pagerank, HITS, centralnosti temeljene na međupovezanosti i centralnosti temeljene na bliskosti. Rezultati analize mreže sljedbenika su potvrdili navode iz literature, dok se za druge dvije mreže pokazalo da su premalene odnosno da se je potrebno sakupljati podatke dulje vrijeme kako bi se došlo do statistički značajnog udjela podataka.

Abstract (english)

Lately, more and more data is getting stored and analyzing it brings with itself new challenges because of its volume, velocity and variety. Twitter is a social network that is a good example of a Big Data source. Through Twitter it is very easy to monitor the information propagation process and that ease greatly improves the ability to study social influence. In this context, social influence can be defined as the ability of a user to influence other users in any way. In this thesis Twitter data was gathered about tweets related to graphics cards as well as data about their authors. That data was used to generate networks of followers, mentions and retweets which were passed to the Pagerank, HITS, closeness and betweenness centrality algorithms. The results of the follower network analysis confirmed what was stated in many of the related articles, while the analysis of the other two networks failed to yield any meaningful results since the amount of gathered data was too small and should be increased if the data sample is to become statistically significant.

• Keywords: veliki skupovi podataka; društveni utjecaj; twitter; analiza veza
• Keywords (english): big data; social influence; twitter; link analysis
• Language: croatian
• URN:NBN: urn:nbn:hr:168:292606
• 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: 2018-02-06 10:34:51