User profile for provisioning information and communication services based on user influence

doctoral thesis

• Title (croatian): Korisnički profič za pružanje informacijskih i komunikacijskih usluga zasnovanih na utjecajnosti korisnika
• Author: Vanja Smailović
• 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: 2016, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Electrical Engineering; Telecommunications and Informatics
• Universal decimal classification (UDC): 621.3 - Electrical engineering

Abstract

User profile enables collection, storage and interpretation of user data, which in turn enables analysis and reasoning upon such data. It is an ongoing research challenge to utilize vast amounts of available multi-source, heterogeneous user data with the goal of identifying key, socially influential actors for provisioning information and communication services. Novel approach to solving this challenge includes several steps, as proposed in this thesis. First, a novel user profiling method is proposed in order to efficiently acquire, aggregate and consolidate user data from two data-sources – the telecom-operator's network and Interned-based social networks. Second, a novel user profile model is developed for inference upon such user data. Third, improved algorithms for reasoning upon profile data are proposed, resulting in new knowledge about each user – their social influence. The proposed algorithm improvements in calculating social influence are statistically validated and evaluated through a real-world experiment. Substantial difference in results between the two data-sources is statistically proven, explaining their synergy once utilized together. Finally, out of several algorithm proposals stemming from the Limited Recursive Algorithm (LRA), the variant with Sample-Literature-Optimal Posting Frequency factor (SLOF) demonstrates significant improvement. For example, churn-prevention, prioritizing customer care, digital advertising and disease-tracking algorithms could all benefit from the SLOF algorithm as exemplified in the thesis.

Abstract (croatian)

Korisnički profil omogućava prikupljanje, pohranu i interpretaciju korisničkih podataka, što zauzvrat omogućava analizu i zaključivanje iz tih podataka. Iskoristiti veliku količinu heterogenih podataka iz više izvora, s ciljem određivanja ključnih, društveno utjecajnih pojedinaca u informacijskim i komunikacijskim uslugama, aktualan je istraživački izazov. Nov pristup rješavanju ovog izazova, predložen u disertaciji, sastoji se od nekoliko koraka. Prvo, predložena metoda za profiliranje korisnika omogućuje učinkovito dohvaćanje, agregaciju i konsolidaciju korisničkih podataka iz dva izvora – telekom-operatorske mreže i usluga za društveno umrežavanje. Drugo, razvijen je model korisničkog profila za automatsko zaključivanje iz tih korisničkih podataka. Treće, predložen je nadograđeni algoritam za zaključivanje iz korisničkih podataka koji rezultira novim znanjem o svakom korisniku – društvenom utjecajnošću tog korisnika. Predložene algoritamske nadogradnje u izračunu društvene utjecajnosti statistički su validirane i ocjenjene kroz eksperiment sa stvarnim korisničkim podacima. Značajna razlika u rezultatima iz dva izvora podataka statistički je dokazana te objašnjena njihovom sinergijom kada se koriste istovremeno. Među predloženim algoritamskim nadogradnjama zasnovanim na "Limited Recursive Algorithm" (LRA), algoritam "Sample-Literature-Optimal Posting Frequency factor" (SLOF) pokazuje značajno poboljšanje. Primjerice, SLOF ima primjenu u poboljšanom sprječavanju odljeva pretplatnika, prioretizaciji usluge, digitalnom oglašavanju i algoritamskom praćenju zaraza.

• Keywords: social networking service; user profile; user profiling; social networks; social network analysis; social influence calculation
• Keywords (croatian): usluga društvenog umrežavanja; korisnički profil; korisničko profiliranje; društvene mreže; analiza društvenih mreža; izračun društvene utjecajnosti
• Language: english
• URN:NBN: urn:nbn:hr:168:595948
• 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: 153 str. ; 30 cm.
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2019-04-11 15:49:02