Scholarlink

The use of technology in the field of sports has grown significantly over the last few years to improve athletes’ performance, reduce the risk of injury, and provide real-time feedback on performance. Change of Direction (COD) is an essential athletic movement that significantly impacts the athlete’s body and performance. Automated methods using Machine learning (ML) and Deep Learning (DL) models using different sensor data have been proven to provide quick or real-time feedback. This study evaluates various DL and ML models based on our literature-informed methodology, but instead of using the conventional method of training DL models on raw data, it uses an identical set of input features, extracted from the raw data, on both the ML and DL models being investigated. The study finds comparable performance in classifying COD movement from the best-performing ML and DL models.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Comparison of Machine Learning and Deep Learning Models for Change of Direction Classification Using Engineered Features

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Pranay Jaiswal,
  • Abhishek Kaushik,
  • Fiona Lawless,
  • Tiago Malaquias,
  • Fergal McCaffery

摘要

The use of technology in the field of sports has grown significantly over the last few years to improve athletes’ performance, reduce the risk of injury, and provide real-time feedback on performance. Change of Direction (COD) is an essential athletic movement that significantly impacts the athlete’s body and performance. Automated methods using Machine learning (ML) and Deep Learning (DL) models using different sensor data have been proven to provide quick or real-time feedback. This study evaluates various DL and ML models based on our literature-informed methodology, but instead of using the conventional method of training DL models on raw data, it uses an identical set of input features, extracted from the raw data, on both the ML and DL models being investigated. The study finds comparable performance in classifying COD movement from the best-performing ML and DL models.