Uzorkovanje na kompleksnim mrežama metodom snježne grude

master's thesis

• Title (english): Snowball sampling method on complex networks
• Author: Josip Batur
• Mentor: Hrvoje Štefančić (mentor)
• Mentor: Davor Horvatić (mentor)
• Committee member: Davor Horvatić (predsjednik povjerenstva)
• Committee member: Hrvoje Štefančić (član povjerenstva)
• Committee member: Nikola Poljak (član povjerenstva)
• Committee member: Matko Milin (član povjerenstva)
• Committee member: Danko Radić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Science; (Department of Physics); Zagreb
• Defense date and country: 2019-09-30, Croatia
• Scientific / art field, discipline and subdiscipline: NATURAL SCIENCES; Physics

Abstract

Veliki broj vrlo raznorodnih kompleksnih sustava koje proučavanju fizika, kemija, biologija, sociologija, ekonomija, računalne i druge znanosti mogu se prikazati kao kompleksne mreže. Istraživanje strukture i dinamike kompleksnih mreža, kao i dinamike procesa koji se odvijaju na kompleksnim mrežama predstavlja jedno od najpropulzivnijih interdisciplinarnih područja moderne znanosti. U mnogim situacijama je od važnosti uzorkovanjem istraživati populaciju za koju su odnosi medu jedinkama predstavljeni kompleksnom mrežom. Cilj teme diplomskog rada je istražiti svojstva postupka uzorkovanja metodom snježne grude na kompleksnim mrežama te ispitati da li se i pod kojim uvjetima metoda snježne grude može približiti po svojim svojstvima slučajnom uzorkovanju na kompleksnim mrežama. Naime, metoda snježne grude se standardno klasificira kao neprobabilistička metoda uzorkovanja kod koje iz uzorka nije moguće pouzdano zaključivanje o populaciji. S druge strane, uzorkovanje metodom snježne grude je često jedini realistični pristup uzorkovanju ukoliko je ciljanoj populaciji onemogućen ili otezan pristup, pa tako i probabilističke metode uzorkovanja poput slučajnog uzorkovanja. Specifični cilj diplomskog rada je kvantifikacija razlike slučajnog uzorkovanja i uzorkovanja metodom snježne grude na kompleksnim mrežama. Istraživanje će se provoditi simulacijama uzorkovanja na računalno generiranim kompleksnim mrežama.

Abstract (english)

A large number of various complex systems in physics, chemistry, biology, sociology, economics, computer and other sciences can be represented as complex networks. Studies of structure and dynamics of complex networks, as well as dynamics of processes that take place on complex networks represent one of the most propulsive interdisciplinary areas of modern science. In complex systems, where relations and interactions between its constituent units are represented by complex network, there are many situations when sampling of population is of great importance. The goal of this theses is to examine properties of snowball sampling on complex networks and to find out under which conditions (if any) can snowball sampling assessments get close to random sampling on complex networks. Particularly, snowball sampling method is classified as non-probabilistic, meaning it is not possible to reliably conclude about some property distribution in population through snowball sample. On the other hand, snowball sampling is often the only realistic sampling method, particularly when there is limited or no approach to the population of interest. Specific task of this thesis is quantification of difference between random sampling and snowball sampling on complex networks. Research will be conducted by simulations of sampling on computer generated complex networks.

• Keywords: kompleksne mreže; uzorkovanje; raspodjela; stupanj čvora
• Keywords (english): complex networks; sampling; distribution; node degree
• Language: croatian
• URN:NBN: urn:nbn:hr:217:794593
• Study programme: Title: Physics; specializations in: Research; Course: Research; Study programme type: university; Study level: integrated undergraduate and graduate; Academic / professional title: magistar/magistra fizike (magistar/magistra fizike)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Open access
• Created on: 2020-01-08 09:17:03