Handling limited datasets with neural networks in medical applications: a small-data approach
Conclusion The significance of this work is two-fold: the practical application allows for non-destructive prediction of bone fracture risk, while the novel methodology extends beyond the task considered in this study and provides a general framework for application of regression NNs to medical problems characterised by limited dataset sizes.
Source: Artificial Intelligence in Medicine - Category: Bioinformatics Source Type: research
More News: Bioinformatics | Learning | Orthopaedics | Osteoarthritis | Study | Universities & Medical Training