Utjecaj susjeda na vlastito mišljenje

master's thesis

• Title (english): Influence of Neighbours on One's Own opinion
• Author: Ivan Bestvina
• Mentor: Mile Šikić (mentor)
• Committee member: Mile Šikić (predsjednik povjerenstva)
• Committee member: Zvonko Kostanjčar (član povjerenstva)
• Committee member: Krešimir Križanović (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2017-07-10, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Predviđanje rezultata izbora koristeći informacije o susjedstvu društvene mreže je problem iz teorije kompleksnih mreža koji uključuje napredne tehnike modeliranja i analize kompleksnih mreža. Pod pretpostavkom da društvena mreža pokazuje visok stupanj homofilije, što kod izbora s dvije izborne opcije dovodi do polarizacije i podjele mreže u dva velika klastera, preveli smo ovaj problem u domenu određivanja veličine klastera. Prvi dio rada analizira vezu između susjedstva društvene mreže i opcije za koju centralni čvor glasuje. Sljedeći dio pokazuje kako se ova opcija može izravno i uspješno predvidjeti iz informacija o društvenom susjedstvu. Naposljetku, pokazali smo kako se, modelirajući društvene mreže s proširenim stohastičkim Kroneckerovim grafovima te koristeći Them vs. us estimator, udio glasova u mreži može uspješno odrediti, čak i kada je uzorak iznimno nereprezentativan.

Abstract (english)

Election results prediction using the social network neighbourhood information is a complex network theory problem which includes advanced complex network modelling and complex network analysis techniques. Presuming that social networks display a high degree of homophily, which in binary elections leads to polarization and a split into two giant clusters, we translated this problem into a domain of cluster size inference, for these two clusters. First, we have analysed how the social neighbourhood is related to a person's voting preference. Next, we have shown that the prediction of this preference can be done from the neighbourhood information in a straightforward manner. Finally, we have shown how, by modelling social networks with the expanded stochastic Kronecker graph model and using our novel Them vs. us estimator, network's vote share can successfully be estimated, even from a small and highly biased node sample.

• Keywords: predviđanje rezultata izbora; predviđanje mišljenja; susjedstvo društvene mreže; Kroneckerov model grafova; polarizacija društvene mreže
• Keywords (english): election results prediction; opinion inference; social neighbourhood; Kronecker graph model; social network polarization
• Language: english
• URN:NBN: urn:nbn:hr:168:699442
• 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 19:03:09