Author: Maja Ulaga
Mentor: prof.dr. Bojan Jošt, Fakulteta for sport
Comenthor: prof.dr. Vladislav Rajkovič, Faculty of organizational sciences
Title: STRUCTURE AND INTER - RELATION ANALYSIS OF CHOSEN POTENTIAL DIMENSIONS OF AN ATHLETE SUCCESS MODEL WITH THE »SPORT MANAGER« EXPERT SYSTEM
The purpose of the study was to answer certain chosen fundamental questions of success theory and athlete preparation system, which required the construction of a success model in sport and comparison of some expert system methods (SMMS and DEXi) and machine learning. The subject of the study was oriented into studying a reduced model of potential successfulness in ski-jumps in the morphological, motor, specific-motor, psychological and sociological sub-spaces. The subject sample included 104 Slovene competitors in ski-jumps. Seniors, as well as younger age groups (17-18, 15-16, 13-14, 11-12 and below 10 years of age) were tested. The criterion variable was an assessment of their competitive success (separately for each age category) in the season 1999/2000.
The obtained results showed the usefulness of all the analysed methods, since several artificial intelligence methods have proven to be a practical way to test and improve expert models. Statistical verification of the reduced model of potential successfulness in ski-jumping showed that the model cannot fulfil all the statistical requirements of modelling success in sport, since certain individual variables have a functional relation to the success criterion and certain specific stochastic relations affect the results and should be taken into consideration.
Keywords: success, model, expert modelling, machine learning, ski-jumps