Principal Component Analysis and Variable Selection in Intensive Care Monitoring

摘要

We discuss methods for dimension reduction like the selection of variables and the calculation of principal components in the clinical context of online monitoring in intensive care. Principal component analysis may be advantageous since principal components linearly combine the information in the observed variables in a way that is optimal in a certain sense, but we have to solve problems with the interpretation of the results. Suitable information on this can be obtained applying graphical models^1,2.