Social Network Analysis with Focus on Behavioral Patterns

Kateřina Slaninová

Doctoral thesis

Abstract:

Dizertaˇcn´ı pr´ace pojedn´av´a o anal´yze soci´aln´ıch s´ıt´ı, kter´a je zamˇeˇren´a na vzory chov´an´ı uˇzivatel˚u. Hlavn´ım c´ılem pr´ace je navrhnout novou metodologii pro anal´yzu chov´an´ı uˇzivatel˚u v syst´emech, kter´e zaznamen´avaj´ı aktivity uˇzivatel˚u ve formˇe log˚u. Z´amˇerem bylo nal´ezt skupiny uˇzivatel˚u s podobn´ym chov´an´ım v syst´emu a jejich jasn´a a pˇrehledn´a vizualizace vˇcetnˇe skryt´ych vazeb mezi nimi. Navrhovan´y postup se skl´ad´a z nˇekolika krok˚u, pˇriˇcemˇz kaˇzd´y z nich vyˇzadoval ˇreˇsen´ı specifick´eho probl´emu. Pro jejich ˇreˇsen´ı bylo pouˇzito princip˚u z r˚uzn´ych oblast´ı, jako je anal´yza business proces˚u, anal´yza log˚u, anal´yza komplexn´ıch s´ıt´ı, data mining, umˇel´a inteligence, teorie graf˚u a dalˇs´ı. Postup byl testov´an na nˇekolika experimentech s re´aln´ymi daty ze tˇr´ı oblast´ı: e-learningov´e syst´emy, website anal´yza a modelov´an´ı a simulace. Metodologie m˚uˇze b´yt vyuˇzita v r˚uzn´ych oblastech, napˇr. pro optimalizaci struktury analyzovan´ych syst´em˚u, v recommended syst´emech, pro ˇr´ızen´ı a spr´avu uˇzivatel˚u nebo student˚u, pro marketingov´e ´uˇcely apod.

Abstract:

The doctoral thesis deals with social network analysis with the focus on users’ behavioural patterns. The main goal is to propose a new methodology for analysis of users’ behaviour in the systems, which store users’ activities into log files. The aim is to find user groups with similar behaviour and latent ties between them to clearly present their behaviour in the system. The proposed approach consists of several steps; each of them required a solution of specific problems. The solutions were done using principles from business process analysis, log mining, complex network analysis, data mining, artificial intelligence and graph theory. The proposed methodology has been tested through several experiments performed within three spheres: e-learning systems, website analysis and simulation modelling. The proposed methodology can be used in several spheres for the optimisation of system structures, in recommended systems, for student/user management, for marketing purposes, etc.

Keywords

anal´yza soci´aln´ıch s´ıt´ı, komplexn´ı s´ıtˇe, vzory chov´an´ı, uˇzivatelsk´e profily, chov´an´ı uˇzivatel˚u, shlukov´an´ı, anal´yza log˚u, anal´yza business proces˚

Keywords

social network analysis, complex networks, behavioural patterns, user profiles, users’ behaviour, clustering, log mining, business process mining

Language used: Czech

Date on which the thesis was submitted / produced: 19. 8. 2013

Thesis defence

• Date of defence: 25. 10. 2013
• Supervisor: Václav Snášel
• Reader: Jaroslav Pokorný, Katarzyna Wegrzyn-Wolska, Ivan Zelinka

VŠB - Technical University of Ostrava

Fakulta elektrotechniky a informatiky

Doctoral programme / field:
Informatika, komunikační technologie a aplikovaná matematika / Informatika