Utjecaj parametara sportske dinamike na učinkovitost prediktivnih modela u sportu

doctoral thesis

• Title (english): Influence of sports dynamics parameters on the efficiency of predictive models in sports
• Author: Ana Šarčević; MBZ: 38221
• Mentor: Mihaela Vranić (mentor)
• Committee member: Damir Pintar (predsjednik povjerenstva)
• Committee member: Boris Vrdoljak (član povjerenstva)
• Committee member: Ivan Magdalenić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; (Department of Electrical Engineering Fundamentals and Measurements); Zagreb
• Defense date and country: 2023, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Electrical Engineering
• Universal decimal classification (UDC): 621.3 - Electrical engineering

Abstract

Priroda sporta uvelike utječe na prilagodbu naprednih tehnika koje će se koristiti za predviđanja. Ova doktorska disertacija fokusirana je na skupinu sportova čija se utakmica sastoji od niza nadigravanja i igra se dok se ne ostvari bodovni cilj, a na takve sportove referira se pojmom sportovi s bodovnim ograničenjem. Fokus disertacije se usmjerava na izgradnju prediktivnog modela odbojkaških utakmica, a model se može generalizirati i na ostale sportove koji koriste sličan sustav bodovanja. U modele za predviđanje ishoda odbojkaških utakmica, disertacija uvodi dva parametra sportske dinamike - parametar kratkoročnog zamaha i parametar dugoročnog zamaha. Kroz formulaciju kratkoročnog zamaha, disertacija opisuje kratkoročne iznadprosječne izvedbe igrača nakon određenog motivirajućeg događaja u utakmici. Dugoročni zamah, s druge strane, kombinira povijesne statistike momčadi i simulacijske podatke utakmica čiji se ishodi previđaju. Disertacija nudi matematičke modele za spomenuta dva parametra sportske dinamike koji se temelje na uvjetnoj vjerojatnosti i empirijskom Bayesovom pravilu ažuriranja te nudi hibridni prediktivni model koji objedinjuje navedene parametre, a temelji se na simulaciji Monte Carlo. Znanstveni doprinos ovog rada sastoji se u sljedećem: 1. Formalno definiranje parametara sportske dinamike u sportovima s bodovnim ograničenjem koje se temelji na uvjetnoj vjerojatnosti i empirijskom Bayesovom pravilu ažuriranja. 2. Postupak oblikovanja modela za predviđanje razlike u broju osvojenih bodova suparnika i ukupnog broja dovršenih nadigravanja u utakmicama sportova s bodovnim ograničenjem koji uzima u obzir parametre sportske dinamike, a temelji se na simulaciji Monte Carlo. 3. Analiza utjecaja parametara sportske dinamike na učinkovitost modela za predviđanje razlike u broju osvojenih bodova suparnika i ukupnog broja dovršenih nadigravanja u utakmicama sportova s bodovnim ograničenjem.

Abstract (english)

The nature of sport influences the adaptation of techniques used for prediction. This dissertation focuses on a group of sports characterized by matches consisting of a series of rallies that continue until a specific point goal is achieved. These sports are commonly referred to as sports with a point limit. While the primary focus of this dissertation is on building a predictive model for volleyball matches, the model can be generalized to other sports. The dissertation introduces two parameters that capture the dynamics of these sports: short-term and long-term momentum. Short-term momentum is formulated to describe the above-average performances of teams following specific motivating events during a match. Long-term momentum combines historical team statistics with simulation data from matches whose outcomes are predicted. Mathematical models based on conditional probability and empirical Bayes estimation are proposed to implement these momentums, which are combined through an approach based on Monte Carlo simulation. The scientific contribution consists of the following: 1. A formal definition of sports dynamics parameters for sports with a point limit, based on conditional probability and the empirical Bayes update rule. 2. The development of a predictive model for estimating the difference in the number of points scored by opponents and the total number of completed rallies in matches of sports with a point limit. This model incorporates the sports dynamics parameters and utilizes Monte Carlo simulation. 3. An analysis of the influence of sports dynamics parameters on the effectiveness of the predictive model for estimating the difference in the number of points scored by opponents and the total number of completed rallies in matches of sports with a point limit.

• Keywords: predvidanje ishoda sportskih događdaja; odbojka; jednolika i neovisna distribucija; Markovljevi lanci; Bayesova statistika; Monte Carlo simulacija; psihološki zamah; sportska dinamika.
• Keywords (english): predicting the outcome of sports events; volleyball; uniform and independent distribution; Markov chains; Bayesian statistics; Monte Carlo simulation; psychological momentum; sports dynamics
• Language: croatian
• URN:NBN: urn:nbn:hr:168:593839
• Promotion: 2023
• Study programme: Title: Doctoral study programme "Electrical Engineering and Computing"; Study programme type: university; Study level: postgraduate; Academic / professional title: doktor/doktorica znanosti, po-dručje tehničkih znanosti (doktor/doktorica znanosti, po-dručje tehničkih znanosti)
• Type of resource: Text
• Extent: xi, 112 str. : grafički prikazi ; 31 cm.
• File origin: Born digital
• Access conditions: Open access
• Created on: 2023-09-18 15:04:40