Modeling performance in IRONMAN® 70.3 age group triathletes using a machine learning model (Preprint)

Author:

Thuany Mabliny,Valero DavidORCID,Villiger Elias,Forte PedroORCID,Weiss KatjaORCID,Andrade Marilia,Nikolaidis Pantelis,Cuk Ivan,Knechtle beatORCID

Abstract

BACKGROUND

Individual factors related to performance in age group triathletes competing in different race distances have been explored in scientific literature. However, only a few studies have been conducted using machine learning (ML) predictive models to explore the im-portance of those individual factors.

OBJECTIVE

This study intended to build and analyze machine learning regression models that pre-dict the performance of IRONMAN® 70.3 age group triathletes, considering sex, age, country of origin, and event location as predictive factors.

METHODS

A total of 823,464 finishers´ records (625,398 men and 198,066 women) of IRONMAN® 70.3 age group triathletes from 240 different countries and participating in 197 different events in 183 different locations between 2004 and 2020 were analyzed. The triathletes’ gender, age, country of origin, event location and year, and race finish times were thus obtained and considered for the study. Four different ML regression models were built to predict the triathletes’ race times from their age, gender, country of origin, and race location. The model with the best performance was then selected and further analyzed using model-agnostic interpretability tools to understand which factors would contribute most to the model predictions.

RESULTS

The Random Forest Regressor model obtained the best predictive score. This model's partial dependence plots indicated that men under 30 years, from Switzerland or Denmark, competing in IRONMAN®70.3 Austria/St. Polten, IRONMAN® 70.3 Switzerland, IRONMAN® 70.3 Sunshine Coast, and IRONMAN® 70.3 Busselton presented the best performance.

CONCLUSIONS

Our results prove that ML models can be used to look into the complex, non-linear interactions between the factors that influence performance and gain insights that can help IRONMAN® 70.3 age group triathletes better plan their races.

Publisher

JMIR Publications Inc.

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3