Artificial Intelligence, Machine Learning, Healthcare, Sports Injuries, Imbalanced data
Abstract
This paper aims to evaluate machine learning applications to obtain artificial intelligence (AI) models for sports injury predictions. Using a dataset from NFL's contest as a cornerstone, various data processing techniques are used, such as dimensionality reduction (PCA and feature importance analysis) and data balancing (upweighting and downsizing), to enable the training of AI models. Finally, AI models are evaluated and compared for injury prediction. The results indicate accuracies above 90% in the tests and suggest that AI techniques are promising solutions for sports injury predictions.
