# Machine Learning and Non-Invasive Monitoring Technologies for Training Load Management in Women’s Volleyball: A Scoping Review

## TL;DR

This review explores how non-invasive technologies and machine learning can help manage training loads in women's volleyball to improve performance and reduce injury risks.

## Contribution

The study provides a comprehensive overview of current technologies and ML applications in women’s volleyball training load management, highlighting gaps in research.

## Key findings

- Inertial measurement units and GPS are commonly used for monitoring training loads in women’s volleyball.
- Machine learning models are underutilized and face challenges in validation and interpretability.
- Few studies consider women-specific factors like the menstrual cycle in training load management.

## Abstract

Training load monitoring in women’s volleyball is a challenge for optimizing performance and mitigating injury risk. Non-invasive monitoring technologies and machine learning (ML) can support decision-making, but the evidence remains heterogeneous. This scoping review mapped and integrated the evidence on training load management, fatigue, and performance in women’s volleyball and identified gaps. The PRISMA Extension for Scoping Reviews (PRISMA-ScR) and the Joanna Briggs Institute (JBI) framework were followed. A systematic search was conducted in Scopus, Web of Science, and PubMed, covering January 2020 to September 2025. We included studies in female players at any competitive level, including mixed-sex studies meeting a minimum threshold of female participation, that evaluated external and/or internal load, neuromuscular or perceptual fatigue, and/or performance, using standardized data extraction and narrative/thematic synthesis. Fifty-three studies were included. Inertial measurement units (IMUs), force platforms, heart rate (HR) and heart rate variability (HRV), wellness questionnaires, and global/local positioning systems (GPSs/LPSs) were most prevalent. External-load intensity indicators (e.g., high-intensity jumps and accelerations) were reported as more sensitive to fatigue-related changes than accumulated volume. Machine learning models were less frequent and were mainly applied to multi-source integration and fatigue/readiness prediction, with recurring limitations in external validation and interpretability. Women-specific biological moderators, such as the menstrual cycle, were rarely addressed.