A Machine Learning-Based Model for Comprehensive Assessment and Classification of Cross-Country Skiing Athletes.

Area of Science:

• Sports Science
• Machine Learning
• Data Science

Background:

• Accurate athlete evaluation is crucial for performance improvement in cross-country skiing.
• Current methods may lack the precision to capture nuanced skill differences.
• A data-driven approach can offer a more objective assessment.

Purpose of the Study:

• To develop and validate a novel machine learning model for categorizing cross-country skiers into five distinct skill levels.
• To integrate diverse athlete metrics (physical, psychological, professional skills) for comprehensive assessment.
• To demonstrate the superiority of the proposed model over existing methods and manual evaluation.

Main Methods:

• A hybrid machine learning model combining Multi-Layer Perceptrons (MLP), Attention mechanisms, and Support Vector Machines (SVM).
• MLP models process individual metric categories (physical, psychological, skills) to extract features.
• Attention mechanisms dynamically weight and fuse features, followed by SVM for final classification.

Main Results:

• The model achieved a high F1 score of 0.9306, significantly outperforming baseline methods by over 0.10.
• Demonstrated superior performance compared to traditional manual grading, offering a more scientific evaluation.
• The Attention mechanism proved effective in dynamic feature weighting and fusion for accurate classification.

Conclusions:

• The proposed machine learning model provides a robust and accurate system for cross-country skier skill level evaluation.
• Dynamic feature integration via Attention mechanisms enhances classification precision.
• This approach offers a more objective, scientific, and reliable alternative to manual assessment in sports analytics.