Handling limited datasets with neural networks in medical applications: A small-data approach

Key Points

• This work aims to develop a method for predicting bone fracture risk using neural networks in settings with limited data.
• Application of regression neural networks to predict outcomes.
• Utilization of a novel methodology that addresses challenges of small dataset sizes.
• Demonstrated successful non-destructive prediction of bone fracture risk.
• Framework developed can be applied to various medical problems with limited datasets.