Multivariate analysis techniques: Multiple discriminant analysis and multiple logistic regression. A study in the theoretical background and application of multiple discriminant analysis and multiple logistic regression

摘要

Multivariate data analysis techniques have been widely used to analyze the simultaneous relationships among correlated variables. This study investigates two multivariate data analysis techniques which attempt to explain or predict one or more dependent measures based on the set of predictor variables. Multiple discriminant analysis and multiple logistic regression both attempt to accurately classify individuals or objects into mutually exclusive groups based on several predictor variables.;In addition, utilizing data from Humana Hospital University of Louisville's burn unit and using multiple discriminant analysis along with multiple logistic regression, this study examines the impact of burn injury and burn treatment on the susceptibility to infections. The results provide evidence that the susceptibility to infections can be significantly influenced by predictors such as age, percentage of burn injury, percentage of skin graft loss, eschar removal, number of operations, inhalation injury, and the number of blood transfusions.