Publications by authors named "Charbel Marche"
Article Synopsis
- The study aimed to compare the accuracy of feed-forward neural networks (FFNN) and neural machine translation (NMT) models in estimating injury severity directly and indirectly using AIS codes.
- Results indicated that indirect estimation via NMT was the most accurate method for predicting high injury severity (ISS ≥ 16), outperforming direct estimation approaches.
- While training times were similar across all models, FFNN models demonstrated significantly faster testing times compared to NMT models.
The injury severity classifications generated from the Abbreviated Injury Scale (AIS) provide information that allows for standardized comparisons in the field of trauma injury research. However, the majority of injuries are coded in International Classification of Diseases (ICD) and lack this severity information. A system to predict injury severity classifications from ICD codes would be beneficial as manually coding in AIS can be time-intensive or even impossible for some retrospective cases.
View Article and Find Full Text PDF