Running Race Times Prediction and Runner Performances Comparison using a Matrix Factorization Approach

Autor: de Smet, Dimitri, Verleysen, Michel, Francaux, Marc, 5th International Congress on Sport Sciences Research and Technology Support
Přispěvatelé: UCL - SST/ICTM/ELEN - Pôle en ingénierie électrique
Rok vydání: 2017
Předmět:
Zdroj: icSPORTS
DOI: 10.5220/0006499300960101
Popis: This work provides tools based on matrix factorization that can be used to predict athlete running race times based on known race results. This is of interest for athlete preparation, for workout route planning and for race events organization. This work differentiates from previous ones by jointly considering athletes and routes. This work shows how race records can be used to infer knowledge on the users and the races. The same tools can also serve to compare different athlete performances and track athlete level over time. Experiments were conducted on race records of 648 athletes from casual to elite levels. Experiments show that the methodology can be applied to real data and gives relevant insights.
Databáze: OpenAIRE