Rangiranje sportskih ekipa

master's thesis

• Title (english): Ranking of sports teams
• Author: Filip Gojšić
• Mentor: Damir Bakić (mentor)
• Committee member: Damir Bakić (predsjednik povjerenstva)
• Committee member: Boris Širola (član povjerenstva)
• Committee member: Domagoj Vrgoč (član povjerenstva)
• Committee member: Franka Miriam Brückler (član povjerenstva)
• Granter: University of Zagreb; Faculty of Science; (Department of Mathematics); Zagreb
• Defense date and country: 2022-09-27, Croatia
• Scientific / art field, discipline and subdiscipline: NATURAL SCIENCES; Mathematics

Abstract

PageRank je algoritam kreiran za rangiranje internetskih stranica koje su međusobno povezane linkovima. Algoritam je baziran na ideji da što više linkova vodi na određenu stranicu to je stranica bolja odnosno njen rang je veći, no ne vrijede svi linkovi jednako. Linkovi s bolje rangiranih stranica vrijede više od onih s loše rangiranih. Također, link sa stranice koja posjeduje linkove na malo stranica (ili samo jednu) vrijedi više od linka sa stranice koja posjeduje puno linkova. Nadalje, potrebno je uračunati i mogućnost da korisnik ne koristi linkove već direktno ide na proizvoljnu stranicu. Algoritam je moguće izmijeniti tako da umjesto internetskih stranica rangira sportske ekipe; u našem slučaju konkretno timove u NFL-u. Zaključili smo, bazirano na podacima iz 2021. sezone, da je PageRank algoritam predviđao pobjednike točnije nego kladionice, profesionalni prognozeri te službena stranica lige.

Abstract (english)

PageRank is an algorithm used to ranking web pages which are interconnected by links. The algorithm is based on the idea that the more links lead to a certain page, the better the page is, that is, its ranking is higher, but not all links are equally valuable. Links from higher-ranking sites are worth more than those from poorly-ranked ones. Also, a link from a page that has links to few pages (or just one) is worth more than a link from a page that contains a lot of links. Furthermore, it is necessary to take into account the possibility that the user does not use link but goes directly to an arbitrary page. The algorithm can be changed so that it ranks sports teams instead of websites; in our case NFL teams. We concluded, based on data from the 2021 season, that the PageRank algorithm predicted winners more accurately than bookmakers, professional analysts and the league’s official website.

• Keywords: PageRank; rangiranje internetskih stranica; timovi u NFL-u; predviđanje pobjednika
• Keywords (english): PageRank; ranking of websites; NFL teams; predicted winners
• Language: croatian
• URN:NBN: urn:nbn:hr:217:192838
• Study programme: Title: Finance and Business Mathematics; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra matematike (magistar/magistra matematike)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Open access
• Created on: 2022-10-26 10:52:50