One Size Doesn't Fit All: Supervised Machine Learning Classification in Athlete-Monitoring

Abstract

Athlete movement data is integral for optimizing athlete-performance and can lead to reduced fatigue and in turn can mitigate injury risk. There is a substantial amount of scientific literature which investigates the ability of computervision and inertial sensor technologies to classify sport-specific movements. The coupling of automatic sport action labelling and athlete-monitoring data can significantly enhance athlete work-load monitoring. Two recent systematic reviews of the literature, pertinent to sport-specific movement classification, revealed that the majority of journal articles use athlete-dependent classification model training and evaluation methods. These methods can significantly enhance model classification performance, particularly with movements which have high inter-athlete technique variation. This is because it enables models to learn features distinctive to all athletes during training. This manuscript details the training and evaluation of supervised machine learning models to automatically classify running surface (athletics track, hard sand and soft sand) using features extracted from an upper-back inertial measurement unit sensor. Possible classification performance enhancement is demonstrated by comparing athlete independent and athlete dependent supervised machine learning methods. Using athlete dependent methods significantly increased the classification performance in terms of weighted average precision, recall, F1-score and accuracy (p < 0.05).

Label-Free Neuropeptide Detection beyond the Debye Length Limit

“…The collected AC-mode sensor data was analyzed with machine learning (ML) classification in an attempt to classify NPY concentrations based on observed currents. Machine learning has revolutionized the field of data analysis and has allowed for the interpretation and classification of complex data without human intervention and has been applied in various fields, including the analysis of sensor data and spectra − and human performance. , The results of classification with only three categories are shown in the SI with a maximum average accuracy of 74.38 ± 7.14% seen for an Ada boost classifier. This result could be useful for NPY screening and eventually screening for human performance.…”