Sustav za profiliranje sportaša zasnovan na znanosti o podacima: studijski slučaj igrača košarke

master's thesis

• Title (english): System for profiling of athletes based on data science: case study of basketball players
• Author: Matija Šekrst
• Mentor: Vedran Podobnik (mentor)
• Committee member: Vedran Podobnik (predsjednik povjerenstva)
• Committee member: Ognjen Dobrijević (član povjerenstva)
• Committee member: Marina Bagić Babac (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2018-07-13, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Prikupljanje i obrada podataka su ključni za razvoj svake grane, pa tako i sporta. Velika količina podataka omogućuje donošenje zaključaka na svim razinama, od trenera momčadi, pa sve do navijača. Fokus ovog rada je na predviđanju ishoda košarkaških utakmica jer mogućnost da se predvidi pobjednik susreta otvara mnogobrojne mogućnosti. Rad pokazuje da je naivni Bayesov model najbolji za predviđanje pobjednika košarkaških utakmica, da su predviđanja svih korištenih modela konzistentna po točnosti i AUC vrijednosti i da je ekonomska funkcija troška nestabilna, pa se ne može donijeti zaključak o isplativosti klađenja na predviđanja.

Abstract (english)

Data acquisition and analysis has become key factor in every industry, including the sport industry. New conclusions are drawn based on large data quantity on every level, from team managers to supporters. This thesis focuses on predictions of basketball games because ability to predict winner of a match could be beneficial. Thesis also shows that Naive Bayes model has the best performance when predicting winner of a basketball game. It also shows that model accuracy and AUC values are consistent and economical cost function is unstable so no valid conclusions can be drawn about betting profitability.

• Keywords: Znanost o podacima; NBA; Node.js; PostgreSQL; RapidMiner; naivni Bayes; nasumična šuma; generalizirani linearni model; duboko učenje; predviđanje rezultata; AUC; ekonomska funkcija troška; profitabilnost klađenja
• Keywords (english): Data science; NBA; Node.js; PostgreSQL; RapidMiner; Naive Bayes; Random Forest; Generalized Linear Model; Deep Learning; results prediction; AUC; economical cost function; betting profitability
• Language: croatian
• URN:NBN: urn:nbn:hr:168:464574
• Study programme: Title: Computing; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2019-03-14 01:11:55