Prediction of Marathon Performance using Artificial
                    Intelligence

Lerebourg, Lucie; Saboul, Damien; Clémençon, Michel; Coquart, Jérémy

doi:10.1055/a-1993-2371

Cited by 7 publications

(3 citation statements)

References 56 publications

(148 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In a study that utilized variables such as underlying 10-kilometer performance, BMI, age, and gender, the ANN and KNN models were used to predict marathon performance. The results showed that the correlation coe cients of both the ANN and KNN models reached 0.9, and the prediction accuracies were above 94% (with KNN outperforming ANN with an average absolute error of 2.4% compared to 5.6% for ANN), indicating high accuracy 17 . Whether using mathematical models or AI-based intelligent algorithms for predicting running performance, the results demonstrate high precision.…”

Section: Discussionmentioning

confidence: 99%

Accuracy Study of Smartwatch in Predicting Running Performance Across Various Distances: A Scientific Investigation

Dai,

Chen,

et al. 2023

Preprint

View full text Add to dashboard Cite

BACKGROUND Smartwatches for running are highly prevalent among amateur runners. Their reliability and validity need investigation for accurate predicting running performance. OBJECTIVE This study aims to examine the accuracy of the HUAWEI WATCH GT Runner smartwatch in predicting running performance. METHODS A total of 154 amateur runners, comprising 123 men and 31 women, were recruited. After wearing the HUAWEI WATCH GT Runner for a minimum of six weeks, the runners' actual completion times for 5km, 10km, and half marathon distances were measured, resulting in 288 test instances. The predicted completion times for the same distances as displayed on the watch on the day of the test were simultaneously recorded. RESULTS The actual and predicted performances for the 5km, 10km, and 21.1km distances were highly correlated, with r ≥ 0.95 (P < 0.001), r2 ≥ 0.9 for all three distances, and an error rate between the measured and predicted values of less than 3%, and ICC ≥ 0.9. The Bias ± 95%LoA was − 20.6 ± 46.1 seconds for the 5km, 4.1 ± 299.1 seconds for the 10km, and 143.8 ± 400.4 seconds for the half marathon. CONCLUSIONS This study confirms that the smartwatch exhibits high precision in predicting 5km, 10km, and half marathon performances, with an accuracy exceeding 97%. The performance prediction feature of the smartwatch can effectively guide amateur runners in setting reasonable competition goals and preparing for races.

show abstract

Section: Discussionmentioning

confidence: 99%

Accuracy Study of Smartwatch in Predicting Running Performance Across Various Distances: A Scientific Investigation

Dai,

Chen,

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…Previous studies have employed linear regression to formulate prediction equations, prompting us to explore this approach in our study [ 14 , 16 , 32 , 33 ]. Additionally, a recent study employed artificial intelligence methodologies to predict marathon completion times [ 34 ]. We used our marathon epidemiological survey data to make a prediction of half-marathon performance for male recreational half-marathon runners in a feasible, applicable manner.…”

Section: Discussionmentioning

confidence: 99%

Prediction of half-marathon performance of male recreational marathon runners using nomogram

Shu,

Wang,

Zhou

et al. 2024

BMC Sports Sci Med Rehabil

View full text Add to dashboard Cite

Background Long-distance running is a popular competitive sport. We performed the current research as to develop an easily accessible and applicable model to predict half-marathon performance in male recreational half-marathon runners by nomogram. Methods Male recreational half-marathon runners in Zhejiang Province, China were recruited. A set of literature-based and panel-reviewed questionnaires were used to assess the epidemiological conditions of the recruited runners. Descriptive and binary regression analyses were done for the profiling and identification of predictors related to higher half-marathon performance (completing time ≤ 105 min). Participants were assigned to the training set (n = 141) and the testing set (n = 61) randomly. A nomogram was used to visually predict the half-marathon performance, and the receiver operating characteristic (ROC) was used to evaluate the predictive ability of the nomogram. Results A total of 202 participants (median age: 49 years; higher half-marathon performance: 33.7%) were included. After multivariate analysis, three variables remained as significant predictors: longer monthly running distance [adjusted odds ratio (AOR) = 0.992, 95% confidence interval (CI): 0.988 to 0.996, p < 0.001], faster mean training pace (AOR = 2.151, 95% CI: 1.275 to 3.630, p < 0.001), and better sleep quality [the Pittsburgh Sleep Quality Index (PSQI), AOR = 2.390, 95% CI: 1.164 to 4.907, p = 0.018]. The AUC of the training and testing sets in nomogram were 0.750 and 0.743, respectively. Further ternary and linear regression analyses corroborated the primary findings. Conclusions This study developed a nomogram with good potential to predict the half-marathon performance of recreational runners. Our results suggest that longer monthly running distance, faster mean training pace and better sleep quality notably contribute to better half-marathon performance.

show abstract

“…The current research in this area focuses on evaluating the effectiveness of different digital health technologies in objectively measuring physical activity and sedentary behavior (e.g., accelerometers, and wristband devices). Researchers are also investigating the application of digital health interventions to improve individual physical activity [45] and sedentary behavior [46,47], exploring the relationship between digital health technologies and exercise behavior [48,49], and studying the correlation between digital health and exercise performance [50,51]. Prominent digital health technologies used in this area include virtual reality, social media, and wearable devices such as activity monitors, accelerometers, and motion sensors.…”

Section: Discussionmentioning

confidence: 99%

Exploring the Landscape of Digital Health in Sport Science: A Bibliometric Analysis of Global Research Trends and Future Directions (2010-2023)

Wang¹,

Li²,

Wang³

et al. 2023

Preprint

View full text Add to dashboard Cite

Digital health has gained widespread adoption across various disciplines. Within the field of sport science, there is growing interest in digital health, leading to a surge in research activities. The objective of this study is to conduct a comprehensive bibliometric and visualization analysis, tracing the evolutionary trajectory, identifying research hotspots, and uncovering emerging trends in digital health within the domain of sport science. By doing so, this research aims to offer scholars a systematic and comprehensive understanding of the dynamic landscape in this field. To achieve this, the Web of Science core collection was chosen as the primary database. A thorough literature search was performed, focusing on digital health and sport science, resulting in the screening of 1870 English-language documents. The findings revealed a consistent upward growth trend in the annual publication count from 2010 to 2023. Notably, the United States, specifically the Commonwealth of Pennsylvania System of Higher Education, and John B. Cronin, emerged as the most prolific contributors in terms of publications. Collaborative efforts among countries, institutions, and authors were also observed. The research hotspots identified encompassed several areas, including the evaluation of digital health technologies' effectiveness, exploration of the relationship between digital health technologies and physical activity/exercise performance, and the use of digital health technologies in assessing and rehabilitating sports injuries. Among the prominent digital health technologies employed in these studies were virtual reality, social media, and wearable devices. The predominant types of studies utilized meta-analysis, randomized controlled trials, and cross-sectional studies. Looking towards the future, the study identified several key research trends. These include the exploration of telemedicine, artificial intelligence-based machine learning, therapeutic interventions, and physical therapy within the realm of digital health. These emerging areas signify promising directions for future investigation and development in the field of digital health within sport science.

show abstract

Prediction of Marathon Performance using Artificial Intelligence

Cited by 7 publications

References 56 publications

Accuracy Study of Smartwatch in Predicting Running Performance Across Various Distances: A Scientific Investigation

Accuracy Study of Smartwatch in Predicting Running Performance Across Various Distances: A Scientific Investigation

Prediction of half-marathon performance of male recreational marathon runners using nomogram

Exploring the Landscape of Digital Health in Sport Science: A Bibliometric Analysis of Global Research Trends and Future Directions (2010-2023)

Contact Info

Product

Resources

About