Various demographic and medical factors have been linked with mortality after suffering from traumatic injuries such as age and post-injury disability. A considerable amount of literature has been published on the building of trauma prediction models. However, few analyse the features selection criteria. Patient records comprise a large amount of data and numerous variables, and some are more important than others. Highlighting the most influential variables and their correlations would assist in the better use of it. The intention of this study is to clarify several aspects of demographic and medical factors that could affect the outcome of trauma in order to exhibit the interaction between these factors and to represent their relationships. In addition, the aim is to use ranking and feature weights to select the features that increase accuracy and lead to better results.
展开▼