Otkrivanje zajednica u društvenim mrežama zasnovano na izračunu povjerenja među korisnicima

doctoral thesis

• Title (english): Community detection in social networks based on calculation of trust between users
• Author: Darko Štriga
• Mentor: Vedran Podobnik (mentor)
• Committee member: Vedran Podobnik (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; (Department of Telecommunications); Zagreb
• Defense date and country: 2018, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing; Information Systems
• Universal decimal classification (UDC): 654 - Telecommunication and telecontrol

Abstract

Proučavanje istraživačkog pitanja "otkrivanje zajednica u društvenim mrežama zasnovano na određivanju povjerenja između korisnika" provedeno je nad skupom podataka kreiranim temeljem analize više od 250 tisuća korisnika društvene mreže Facebook. Za potrebe rješavanja istraživačkog pitanja predložena je metoda SmartSocial Trust, a uz postojeće mjere evaluacije razvijena je nova mjera koeficijent SmartSocial Trust. Dobiveni rezultati uspoređeni su s teoretskim matematičkim modelom metode za klasifikaciju korisnika društvene mreže u društvene zajednice i referentnom state-of-the-art metodom iTrust. Prema svim parametrima evaluacije metoda SmartSocial Trust pokazala se kao najbolja. Ostvaren je sljedeći izvorni znanstveni doprinos: (i) model zajednica u društvenim mrežama zasnovan na povjerenju među korisnicima izvedenom iz atributa korisničkih profila i strukturnih svojstava društvenog grafa; (ii) metoda otkrivanja zajednica u društvenim mrežama koja obuhvaća postupke transformacije internetske društvene mreže u društveni graf i algoritam za izračun povjerenja u dobivenom društvenom grafu; (iii) laboratorijski prototip za otkrivanje zajednica u društvenim mrežama verificiran na referentnom skupu podataka „SmartSocial Trust“ stvorenom provođenjem društvenog eksperimenta u internetskom okružju.

Abstract (english)

The research question "community detection in social networks based on a calculation of trust between users" was studied on the dataset created through the analysis of more than 250 thousand of Facebook users. The SmartSocial Trust method was proposed for solving the described research question. Further, in addition to the existing evaluation methods, a new measure, called the SmartSocial Trust coefficient, was developed. The achieved results were compared with a mathematical model for classification of social network users in communities based on the random classification and state-of-the-art reference iTrust method. Among the analysed methods, the SmartSocial Trust method proved to be the best according to all evaluation parameters. The original scientific contribution has been achieved as follows: (i) social community model based on a calculation of trust between online social network users, where trust is derived from user profile attributes and structural features of the social graph; (ii) social community detection method, which includes online social network transformation into social graph and the algorithm for calculating social trust based on the social graph; (iii) laboratory prototype for community detection in online social networks verified on the reference dataset “SmartSocial Trust” created by conducting a social experiment in the Internet environment.

• Keywords: analiza društvenih mreža; znanost o podacima; otkrivanje zajednica; povjerenje između korisnika; Facebook; SmartSocial.
• Keywords (english): social network analysis; data science; community detection; trust between users; Facebook; SmartSocial.
• Language: croatian
• URN:NBN: urn:nbn:hr:168:934272
• Study programme: Title: Electrical Engineering and Computing; Study programme type: university; Study level: postgraduate; Academic / professional title: Doktor znanosti elektrotehnike i računarstva (Doktor znanosti elektrotehnike i računarstva)
• Type of resource: Text
• Extent: 150 str. ; 30 cm.
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2019-04-03 14:26:41