DeepSeek-AI-enhanced virtual reality training for mass casualty management: Leveraging machine learning for personalized instructional optimization.
<h4>Objective</h4>This study aimed to evaluate the effectiveness of a virtual reality (VR) training system for mass casualty management, integrating artificial intelligence (AI) and machine learning (ML) to analyze trainee performance and error patterns. The goal was to identify key pred...
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Public Library of Science (PLoS)
2025-01-01
|
| Series: | PLoS ONE |
| Online Access: | https://doi.org/10.1371/journal.pone.0321352 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | <h4>Objective</h4>This study aimed to evaluate the effectiveness of a virtual reality (VR) training system for mass casualty management, integrating artificial intelligence (AI) and machine learning (ML) to analyze trainee performance and error patterns. The goal was to identify key predictors of performance, generate personalized feedback, and provide actionable recommendations for optimizing VR-based medical training.<h4>Materials and methods</h4>A total of 196 medical professionals participated in a 1-hour VR training session, followed by a 20-question assessment and a post-training evaluation survey. The DeepSeek AI framework was employed to analyze the data, utilizing clustering analysis, principal component analysis (PCA), and random forest models. Descriptive statistics, error rates, and correlation analyses were performed using R software (version 4.1.2). Machine learning models were trained to predict performance outcomes, and feature importance was assessed using the Gini index. Personalized feedback reports were generated based on clustering and error analysis results.<h4>Results</h4>The study identified three distinct trainee clusters, with the highest-performing group excelling in Trauma Assessment and Clinical Case Analysis. However, high error rates were observed in Clinical Case Analysis (69.4%) and Trauma Assessment (67.3%), indicating areas for targeted improvement. Machine learning models highlighted replacing traditional teaching methods (IncNodePurity = 25.76) and stimulating learning interest (IncNodePurity = 13.08) as the most critical factors influencing learning outcomes. AI-driven feedback provided actionable recommendations, such as redesigning complex scenarios and enhancing system usability.<h4>Conclusions</h4>This study demonstrates the potential of integrating AI with VR training to create a more personalized and effective learning experience for medical professionals. The findings underscore the importance of adaptive, data-driven approaches in medical education, particularly in high-stakes environments such as emergency medicine. Future research should explore hybrid training models and incorporate physiological data to further enhance the efficacy of VR-based training systems. |
|---|---|
| ISSN: | 1932-6203 |