University of Twente Student Theses
Monitoring training load and identifying fatigue in young elite speed skaters using machine learning methods
Bolten, Marjolein (2024) Monitoring training load and identifying fatigue in young elite speed skaters using machine learning methods.
|
PDF
5MB |
| Abstract: | Monitoring training load is crucial for enhancing sports performance, as excessive load can lead to fatigue accumulation and decreased performance. Research has extensively investigated sports monitoring techniques, including measuring external and internal loads. Understanding fatigue states helps coaches prevent nonfunctional overreaching and optimize training for improved performance. Continuous monitoring, such as using non-invasive maximal effort tests, is essential to detect declines in performance and adjust training accordingly. In this study, a dataset is collected of young elite speed skaters consisting of data from morning questionnaires, daily jump tests, Wingates tests, and information about training sessions. Despite the amount of research already done in the direction of fatigue and sports performance some connections are still not described in full detail. To get a better understanding of these connections, three different methods are examined on this dataset in relation to fatigue. As a first method, the importance of input variables is determined using a classification decision tree method. Using statistical tests the data of Wingate tests is analyzed and lastly, different Long Short-Term Memory (LSTM) models are tested for their ability to predict resting heart rate (HR) data. Monitoring daily jump height and using a wellness questionnaire did not effectively identify fatigue in young elite speed skaters. Similarly, the Wingate test, conducted three times in five weeks, failed to serve as a reliable measure of fatigue due to inconclusive results influenced by other factors. However, a univariate LSTM model showed promise in predicting daily resting HR data, with an average Root-Mean-Square Error (RMSE) of 1.5 beats per minute. Before this model can be used in a practical situation, further research is needed to improve the performance of the LSTM model. As a practical application, this model can allow for the detection of abnormal HR patterns indicative of fatigue. Consequently, a combination of monitoring internal and external loads, along with predictive resting HR data using LSTM models, offers a possible viable approach to identifying fatigue in young elite speed skaters. |
| Item Type: | Essay (Master) |
| Faculty: | EEMCS: Electrical Engineering, Mathematics and Computer Science |
| Subject: | 54 computer science |
| Programme: | Computer Science MSc (60300) |
| Link to this item: | https://purl.utwente.nl/essays/98547 |
| Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Show download statistics for this publication
Show download statistics for this publicationRepository Staff Only: item control page